minor...

modules/deeplearning/icing_cnn.py

@@ -150,8 +150,6 @@ class IcingIntensityNN:
         self.test_recall = None
         self.test_precision = None
         self.test_confusion_matrix = None
-        self.test_f1score = None
-        self.test_mcc = None
         self.test_true_pos = None
         self.test_true_neg = None
         self.test_false_pos = None
@@ -485,8 +483,6 @@ class IcingIntensityNN:
             self.test_auc = tf.keras.metrics.AUC(name='test_auc')
             self.test_recall = tf.keras.metrics.Recall(name='test_recall')
             self.test_precision = tf.keras.metrics.Precision(name='test_precision')
-            self.test_f1score = tfa.metrics.F1Score(2, threshold=0.5, name='test_f1score')
-            self.test_mcc = tfa.metrics.MatthewsCorrelationCoefficient(2, name='test_mcc')
             self.test_true_neg = tf.keras.metrics.TrueNegatives(name='test_true_neg')
             self.test_true_pos = tf.keras.metrics.TruePositives(name='test_true_pos')
             self.test_false_neg = tf.keras.metrics.FalseNegatives(name='test_false_neg')
@@ -546,8 +542,6 @@ class IcingIntensityNN:
             self.test_true_pos(labels, pred)
             self.test_false_neg(labels, pred)
             self.test_false_pos(labels, pred)
-            self.test_f1score(labels, pred)
-            self.test_mcc(labels, pred)
 
     def predict(self, mini_batch):
         inputs = [mini_batch[0]]
@@ -574,15 +568,23 @@ class IcingIntensityNN:
         self.test_auc.reset_states()
         self.test_recall.reset_states()
         self.test_precision.reset_states()
-        self.test_f1score.reset_states()
-        self.test_mcc.reset_states()
         self.test_true_neg.reset_states()
         self.test_true_pos.reset_states()
         self.test_false_neg.reset_states()
         self.test_false_pos.reset_states()
 
-    def do_metrics(self):
-        pass
+    def get_metrics(self):
+        recall = self.test_recall.result()
+        precsn = self.test_precision.result()
+        f1 = 2 * (precsn * recall) / (precsn + recall)
+
+        tn = self.test_true_neg.result()
+        tp = self.test_true_pos.result()
+        fn = self.test_false_neg.result()
+        fp = self.test_false_pos.result()
+
+        mcc = ((tp * tn) - (fp * fn)) / np.sqrt((tp + fp) * (tp + fn) * (tn + fp) * (tn + fn))
+        return f1, mcc
 
     def do_training(self, ckpt_dir=None):
 
@@ -631,16 +633,18 @@ class IcingIntensityNN:
                             for mini_batch_test in tst_ds:
                                 self.test_step(mini_batch_test)
 
-                        recall = self.test_recall.result()
-                        precsn = self.test_precision.result()
-                        f1 = 2 * (precsn * recall) / (precsn + recall)
-
-                        tn = self.test_true_neg.result()
-                        tp = self.test_true_pos.result()
-                        fn = self.test_false_neg.result()
-                        fp = self.test_false_pos.result()
+                        # recall = self.test_recall.result()
+                        # precsn = self.test_precision.result()
+                        # f1 = 2 * (precsn * recall) / (precsn + recall)
+                        #
+                        # tn = self.test_true_neg.result()
+                        # tp = self.test_true_pos.result()
+                        # fn = self.test_false_neg.result()
+                        # fp = self.test_false_pos.result()
+                        #
+                        # mcc = ((tp * tn) - (fp * fn)) / np.sqrt((tp + fp) * (tp + fn) * (tn + fp) * (tn + fn))
 
-                        mcc = ((tp * tn) - (fp * fn)) / np.sqrt((tp + fp) * (tp + fn) * (tn + fp) * (tn + fn))
+                        f1, mcc = self.get_metrics()
 
                         with self.writer_valid.as_default():
                             tf.summary.scalar('loss_val', self.test_loss.result(), step=step)
@@ -673,20 +677,21 @@ class IcingIntensityNN:
                 for mini_batch in ds:
                     self.test_step(mini_batch)
 
-            recall = self.test_recall.result().numpy()
-            precsn = self.test_precision.result().numpy()
-            f1 = 2 * (precsn * recall) / (precsn + recall)
-
-            tn = self.test_true_neg.result().numpy()
-            tp = self.test_true_pos.result().numpy()
-            fn = self.test_false_neg.result().numpy()
-            fp = self.test_false_pos.result().numpy()
-
-            mcc = ((tp*tn) - (fp*fn))/np.sqrt((tp+fp)*(tp+fn)*(tn+fp)*(tn+fn))
+            # recall = self.test_recall.result().numpy()
+            # precsn = self.test_precision.result().numpy()
+            # f1 = 2 * (precsn * recall) / (precsn + recall)
+            #
+            # tn = self.test_true_neg.result().numpy()
+            # tp = self.test_true_pos.result().numpy()
+            # fn = self.test_false_neg.result().numpy()
+            # fp = self.test_false_pos.result().numpy()
+            #
+            # mcc = ((tp*tn)-(fp*fn))/np.sqrt((tp+fp)*(tp+fn)*(tn+fp)*(tn+fn))
+            f1, mcc = self.get_metrics()
 
             print('loss, acc, recall, precision, auc, f1, mcc: ', self.test_loss.result().numpy(), self.test_accuracy.result().numpy(),
-                  recall, precsn, self.test_auc.result().numpy(), f1, self.test_f1score.result().numpy(), mcc, self.test_mcc.result().numpy())
-            print('--------------------------------------------------')
+                  self.test_recall.result().numpy(), self.test_precision.result().numpy(), self.test_auc.result().numpy(), f1.numpy(), mcc.numpy())
+            print('------------------------------------------------------')
             ckpt_manager.save()
 
             if self.DISK_CACHE and epoch == 0: