diff --git a/modules/deeplearning/cnn_cld_frac_mod_res.py b/modules/deeplearning/cnn_cld_frac_mod_res.py
index eb12f2313789f3eccd074f0d6ea79fb03bf49ffc..c754bbbafb87476d25d5bc53e09754a90e4f90ae 100644
--- a/modules/deeplearning/cnn_cld_frac_mod_res.py
+++ b/modules/deeplearning/cnn_cld_frac_mod_res.py
@@ -603,6 +603,7 @@ class SRCNN:
         self.test_accuracy(labels, pred)
 
     # @tf.function(input_signature=[tf.TensorSpec(None, tf.float32), tf.TensorSpec(None, tf.float32)])
+    # decorator commented out because pred.numpy(): pred not evaluated yet.
     def predict(self, inputs, labels):
         labels = tf.squeeze(labels)
         pred = self.model([inputs], training=False)