Improved code and organization (bug fixed)

Examples/MNIST.py

@@ -36,6 +36,6 @@
     batch_size = 500
     sparse_categorical = 0
     n_epochs = [10, 500, 10]  ## DNN--RNN-CNN
-    Random_Deep = [1, 0, 1]  ## DNN--RNN-CNN
+    Random_Deep = [0, 0, 1]  ## DNN--RNN-CNN
     RMDL.Image_Classification(X_train, y_train, X_test, y_test, batch_size, shape, sparse_categorical, Random_Deep,
                             n_epochs)

RMDL/BuildModel.py

@@ -23,8 +23,8 @@
 from keras.layers import Dense, Flatten
 from keras.layers import Conv1D,MaxPooling2D, \
     MaxPooling1D, Embedding, Merge, Dropout,\
-    GRU,TimeDistributed,Convolution2D,\
-    Activation,Convolution3D,GlobalAveragePooling3D,LSTM
+    GRU,TimeDistributed,Conv2D,\
+    Activation,Conv3D,GlobalAveragePooling3D,LSTM
 from keras import backend as K
 from keras.models import Model
 from keras.layers import Input
@@ -149,14 +149,14 @@ def Image_model_CNN(num_classes,shape):
     Layers = list(range(1, 4))
     Layer = random.choice(Layers)
     Filter = random.choice(values)
-    model.add(Convolution2D(Filter, 3, 3, border_mode='same', input_shape=shape))
+    model.add(Conv2D(Filter, (3, 3), padding='same', input_shape=shape))
     model.add(Activation('relu'))
-    model.add(Convolution2D(Filter, 3, 3))
+    model.add(Conv2D(Filter, (3, 3)))
     model.add(Activation('relu'))
 
     for i in range(0,Layer):
         Filter = random.choice(values)
-        model.add(Convolution2D(Filter, 3, 3,border_mode='same'))
+        model.add(Conv2D(Filter, (3, 3),padding='same'))
         model.add(Activation('relu'))
         model.add(MaxPooling2D(pool_size=(2, 2)))
         model.add(Dropout(0.5))
@@ -180,13 +180,13 @@ def Image_3D_model_CNN(num_classes,shape,kernel_size=(3,3)):
     Layer = 3
     Filter = random.choice(values)
     print(shape)
-    model.add(Convolution3D(Filter,3,3,1,border_mode='same', input_shape=shape))
+    model.add(Conv2D(Filter,(3,3),1,padding='same', input_shape=shape))
     model.add(Activation('relu'))
-    model.add(Convolution3D(Filter,3, 3,1, border_mode='same', subsample=(2, 2,1)))
+    model.add(Conv2D(Filter,(3, 3),1, padding='same', subsample=(2, 2,1)))
     model.add(Dropout(0.25))
     for i in range(0,Layer):
         Filter = random.choice(values)
-        model.add(Convolution3D(Filter, 3,3,1,subsample = (2,2,1)))
+        model.add(Conv3D(Filter,(3,3,1),subsample = (2,2,1)))
         model.add(Activation('relu'))
     model.add(Dense(512,activation='relu'))
     model.add(GlobalAveragePooling3D())

RMDL/Global.py

@@ -19,26 +19,31 @@
 import os
 import os.path
 global GloVe_DIR
-from pathlib import Path
 from Download import Download_Glove as GloVe
 
 
-GloVe_DIR = GloVe.download_and_extract(data='Twitter')
-GloVe_file = "glove.6B.100d.txt"
-GloVe_DIR = os.path.join(GloVe_DIR, GloVe_file)
+GloVe_DIR = ""
+GloVe_DIR = ""
+
+
 
 
-if not os.path.isfile(GloVe_DIR):
-    print("Could not find %s Set GloVe Directory in Global.py ",GloVe)
-    exit()
 
 
 MAX_SEQUENCE_LENGTH = 500
 MAX_NB_WORDS = 75000
 EMBEDDING_DIM = 100  # it could be 50, 100, and 300 please download GLOVE https://nlp.stanford.edu/projects/glove/ and set address of directory in Text_Data_load.py
 
-def setup():
+def setup(text=False):
     np.set_printoptions(threshold=np.inf)
     np.random.seed(7)
     if not os.path.exists(".\weights"):
         os.makedirs(".\weights")
+
+    if text:
+        GloVe_DIR = GloVe.download_and_extract(data='Twitter')
+        GloVe_file = "glove.6B.100d.txt"
+        GloVe_DIR = os.path.join(GloVe_DIR, GloVe_file)
+        if not os.path.isfile(GloVe_DIR):
+            print("Could not find %s Set GloVe Directory in Global.py ", GloVe)
+            exit()

RMDL/RMDL_Image.py

@@ -33,7 +33,7 @@
 
 def Image_Classification(X_train, y_train, X_test, y_test, batch_size, shape, sparse_categorical, Random_Deep,
                             n_epochs):
-    G.setup()
+    G.setup(text=False)
     y_proba = []
 
     score = []

RMDL/RMDL_Text.py

@@ -37,7 +37,7 @@
 def Text_Classification(X_train, y_train, X_test, y_test, batch_size, sparse_categorical, Random_Deep,
                             n_epochs):
 
-    G.setup()
+    G.setup(text=True)
     X_train_tfidf, X_test_tfidf = txt.loadData(X_train, X_test)
     X_train_Embedded, X_test_Embedded, word_index, embeddings_index = txt.loadData_Tokenizer(X_train, X_test)
     del X_train