23-10-21

version 1.0.2
back propagation 설정 가능
=> 초기에 한해서 역전파 1회 실행 가능

fashion_mnist.py

@@ -115,6 +115,7 @@ best_score = pso_mnist.fit(
     empirical_balance=False,
     dispersion=False,
     batch_size=5000,
+    back_propagation=True,
 )
 
 print("Done!")

mnist.py

@@ -117,6 +117,7 @@ best_score = pso_mnist.fit(
     empirical_balance=False,
     dispersion=False,
     batch_size=5000,
+    back_propagation=True,
 )
 
 print("Done!")

pso/__init__.py

@@ -1,7 +1,7 @@
 from .optimizer import Optimizer as optimizer
 from .particle import Particle as particle
 
-__version__ = "1.0.1"
+__version__ = "1.0.2"
 
 __all__ = [
     "optimizer",

pso/optimizer.py

@@ -328,6 +328,7 @@ class Optimizer:
         check_point: int = None,
         batch_size: int = None,
         validate_data: any = None,
+        back_propagation: bool = False,
     ):
         """
         # Args:
@@ -393,20 +394,21 @@ class Optimizer:
         except ValueError as ve:
             sys.exit(ve)
 
-        model_ = keras.models.model_from_json(self.model.to_json())
-        model_.compile(
-            loss=self.loss,
-            optimizer="adam",
-            metrics=["accuracy", "mse"]
-        )
-        model_.fit(x, y, epochs=1, verbose=0)
-        score = model_.evaluate(x, y, verbose=1)
+        if back_propagation:
+            model_ = keras.models.model_from_json(self.model.to_json())
+            model_.compile(
+                loss=self.loss,
+                optimizer="adam",
+                metrics=["accuracy", "mse"]
+            )
+            model_.fit(x, y, epochs=1, verbose=0)
+            score = model_.evaluate(x, y, verbose=1)
 
-        self.g_best_score = score
+            self.g_best_score = score
 
-        self.g_best = model_.get_weights()
+            self.g_best = model_.get_weights()
 
-        del model_
+            del model_
 
         dataset = self._batch_generator_(x, y, batch_size=batch_size)