Make kerasga.predict faster by not cloning the model

predict() cloned the whole model on every call and then used
model.predict(). When predict() runs once per solution inside a fitness
function, this is slow. It now sets the solution as the model weights,
restores the original weights at the end, and calls the model directly
when no batching is asked for. The output is the same and it runs much
faster. model.predict() is still used when batch_size or steps is given.

Author: ahmedfgad

pygad/kerasga/kerasga.py

@@ -79,8 +79,8 @@ def predict(model,
             steps=None):
     """
     Load the given solution as the model's weights and run a forward
-    pass on ``data``. The model is cloned first so the caller's
-    instance is left untouched.
+    pass on ``data``. The model's original weights are restored
+    afterwards, so the model passed by the caller is not changed.
 
     Parameters
     ----------
@@ -105,12 +105,24 @@ def predict(model,
     # Fetch the parameters of the best solution.
     solution_weights = model_weights_as_matrix(model=model,
                                                weights_vector=solution)
-    _model = tensorflow.keras.models.clone_model(model)
-    _model.set_weights(solution_weights)
-    predictions = _model.predict(x=data,
-                                 batch_size=batch_size,
-                                 verbose=verbose,
-                                 steps=steps)
+
+    # Set the solution as the model weights, then put the original weights
+    # back at the end. The model is not cloned because cloning it on every
+    # call is slow when predict() is used inside a fitness function.
+    original_weights = model.get_weights()
+    model.set_weights(solution_weights)
+    try:
+        if batch_size is None and steps is None:
+            # When no batching is asked for, call the model directly. This
+            # is faster than model.predict() when called once per solution.
+            predictions = numpy.array(model(data, training=False))
+        else:
+            predictions = model.predict(x=data,
+                                        batch_size=batch_size,
+                                        verbose=verbose,
+                                        steps=steps)
+    finally:
+        model.set_weights(original_weights)
 
     return predictions