trivial code cleanup

genome.py

@@ -70,17 +70,12 @@ def mutate_one_gene(self):
         gene_to_mutate = random.choice( list(self.all_possible_genes.keys()) )
 
         # And then let's mutate one of the genes.
-        # We need to make sure that this actually creates mutation!!!!!
+        # Make sure that this actually creates mutation
         current_value    = self.geneparam[gene_to_mutate]
         possible_choices = copy.deepcopy(self.all_possible_genes[gene_to_mutate])
-
-        #print(self.all_possible_genes[gene_to_mutate])
-        #print(current_value)
 
         possible_choices.remove(current_value)
 
-        #print(possible_choices)
-
         self.geneparam[gene_to_mutate] = random.choice( possible_choices )
 
         self.update_hash()

idgen.py

@@ -14,7 +14,6 @@ def __init__(self):
 
     def get_next_ID(self):
 
-        #id_to_return = self.currentID
         self.currentID += 1
 
         return self.currentID

main.py

@@ -124,35 +124,35 @@ def main():
         dataset = 'mnist_cnn'
     elif (ds == 3):
         dataset = 'cifar10_mlp'
-    else:
+    elif (ds == 4):
         dataset = 'cifar10_cnn'
+    else:
+        dataset = 'mnist_mlp'
 
     print("***Dataset:", dataset)
 
-    #dataset = 'mnist_mlp' # 'cifar10' 'mnist_mlp' 'mnist_cnn'
-
     if dataset == 'mnist_cnn':
-        generations = 5 # Number of times to evolve the population.
+        generations = 8 # Number of times to evolve the population.
         all_possible_genes = {
-            'nb_neurons': [32, 64, 128],
-            'nb_layers':  [1, 2, 3],
-            'activation': ['relu', 'elu', 'tanh', 'sigmoid'],
+            'nb_neurons': [16, 32, 64, 128],
+            'nb_layers':  [1, 2, 3, 4 ,5],
+            'activation': ['relu', 'elu', 'tanh', 'sigmoid', 'hard_sigmoid','softplus','linear'],
             'optimizer':  ['rmsprop', 'adam', 'sgd', 'adagrad','adadelta', 'adamax', 'nadam']
         }
     elif dataset == 'mnist_mlp':
-        generations = 3 # Number of times to evolve the population.
+        generations = 8 # Number of times to evolve the population.
         all_possible_genes = {
             'nb_neurons': [64, 128], #, 256, 512, 768, 1024],
-            'nb_layers':  [1, 2, 3, 4],
-            'activation': ['relu', 'elu', 'tanh', 'sigmoid'],
+            'nb_layers':  [1, 2, 3, 4, 5],
+            'activation': ['relu', 'elu', 'tanh', 'sigmoid', 'hard_sigmoid','softplus','linear'],
             'optimizer':  ['rmsprop', 'adam', 'sgd', 'adagrad','adadelta', 'adamax', 'nadam']
         }
     elif dataset == 'cifar10_mlp':
         generations = 8 # Number of times to evolve the population.
         all_possible_genes = {
             'nb_neurons': [64, 128, 256, 512, 768, 1024],
-            'nb_layers':  [1, 2, 3, 4],
-            'activation': ['relu', 'elu', 'tanh', 'sigmoid'],
+            'nb_layers':  [1, 2, 3, 4, 5],
+            'activation': ['relu', 'elu', 'tanh', 'sigmoid', 'hard_sigmoid','softplus','linear'],
             'optimizer':  ['rmsprop', 'adam', 'sgd', 'adagrad','adadelta', 'adamax', 'nadam']
         }
     elif dataset == 'cifar10_cnn':
@@ -167,8 +167,8 @@ def main():
         generations = 8 # Number of times to evolve the population.
         all_possible_genes = {
             'nb_neurons': [64, 128, 256, 512, 768, 1024],
-            'nb_layers':  [1, 2, 3, 4],
-            'activation': ['relu', 'elu', 'tanh', 'sigmoid'],
+            'nb_layers':  [1, 2, 3, 4, 5],
+            'activation': ['relu', 'elu', 'tanh', 'sigmoid', 'hard_sigmoid','softplus','linear'],
             'optimizer':  ['rmsprop', 'adam', 'sgd', 'adagrad','adadelta', 'adamax', 'nadam']
         }
 

train.py

@@ -32,7 +32,7 @@ def get_cifar10_mlp():
     # Set defaults.
     nb_classes  = 10 #dataset dependent 
     batch_size  = 64
-    epochs      = 1 #will mostly terminate earlier
+    epochs      = 4
     input_shape = (3072,) #because it's RGB
 
     # Get the data.
@@ -84,8 +84,8 @@ def get_mnist_mlp():
     """Retrieve the MNIST dataset and process the data."""
     # Set defaults.
     nb_classes  = 10 #dataset dependent 
-    batch_size  = 128*12*8 # all we use mlp for is super-crude debugging.
-    epochs      = 1 #will mostly terminate earlier
+    batch_size  = 64
+    epochs      = 4
     input_shape = (784,)
 
     # Get the data.
@@ -227,6 +227,7 @@ def compile_model_cnn(geneparam, nb_classes, input_shape):
     model.add(Dense(nb_classes, activation = 'softmax'))
 
     #BAYESIAN CONVOLUTIONAL NEURAL NETWORKS WITH BERNOULLI APPROXIMATE VARIATIONAL INFERENCE
+    #need to read this paper
 
     model.compile(loss='categorical_crossentropy',
               optimizer=optimizer,