Update to Tensorflow 2

GettingStarted.ipynb

@@ -32,11 +32,11 @@
    ],
    "source": [
     "import numpy as np\n",
-    "from keras.layers import Dense\n",
-    "from keras.models import Sequential\n",
+    "from tensorflow.keras.layers import Dense\n",
+    "from tensorflow.keras.models import Sequential\n",
     "from IPython.display import SVG\n",
-    "from keras.utils import model_to_dot\n",
-    "from tensorflow import set_random_seed"
+    "from tensorflow.keras.utils import model_to_dot\n",
+    "from tensorflow.compat.v1.random import set_random_seed"
    ]
   },
   {

KerasWeightsProcessing/convert_weights.py

@@ -6,12 +6,12 @@
 
 import numpy as np
 import math
-import keras
-import keras.backend as K
-from keras.models import Sequential, Model
-from keras.layers import Dense, Dropout, BatchNormalization
-from keras.layers import Input, Activation
-from keras import optimizers
+from tensorflow import keras
+import tensorflow.keras.backend as K
+from tensorflow.keras.models import Sequential, Model
+from tensorflow.keras.layers import Dense, Dropout, BatchNormalization
+from tensorflow.keras.layers import Input, Activation
+from tensorflow.keras import optimizers
 
 INPUT = ['input']
 ACTIVATIONS = ['relu', 'linear', 'leakyrelu', 'sigmoid']
@@ -127,7 +127,7 @@ def txt_to_h5(weights_file_name, output_file_name=''):
         # if not specified will use path of weights_file with h5 extension
         output_file_name = weights_file_name.replace('.txt', '_converted.h5')
 
-    model.save(output_file_name)
+    model.save(output_file_name, save_format='h5')
 
 def h5_to_txt(weights_file_name, output_file_name=''):
     '''

KerasWeightsProcessing/examples/mnist_keras.py

@@ -2,11 +2,11 @@
 sys.path.append('../')
 from convert_weights import h5_to_txt
 
-import keras
-from keras.datasets import mnist
-from keras.models import Sequential, Model
-from keras.layers import Dense, Input
-from keras.optimizers import RMSprop
+from tensorflow import keras
+from tensorflow.keras.datasets import mnist
+from tensorflow.keras.models import Sequential, Model
+from tensorflow.keras.layers import Dense, Input
+from tensorflow.keras.optimizers import RMSprop
 
 weights_file_name = 'mnist_example.h5'
 txt_file_name     = weights_file_name.replace('h5', 'txt')
@@ -50,6 +50,6 @@
                     verbose=1,
                     validation_data=(x_test, y_test))
 
-model.save(weights_file_name)
+model.save(weights_file_name, save_format='h5')
 
 h5_to_txt(weights_file_name, txt_file_name)

KerasWeightsProcessing/examples/multi_output_model.py

@@ -6,14 +6,14 @@
 import tensorflow as tf
 
 ############## REPRODUCIBILITY ############
-tf.set_random_seed(0)
+tf.compat.v1.set_random_seed(0)
 np.random.seed(0)
 ###########################################
 
-from keras.models import load_model
-from keras.models import Sequential, Model
-from keras.utils.vis_utils import plot_model
-from keras.layers import Dense, BatchNormalization, Input
+from tensorflow.keras.models import load_model
+from tensorflow.keras.models import Sequential, Model
+from tensorflow.compat.v1.keras.utils import plot_model
+from tensorflow.keras.layers import Dense, BatchNormalization, Input
 
 input = x = Input((5,))
 for i in range(3):
@@ -34,7 +34,7 @@
     metrics=['accuracy']
 )
 # SAVE TO FILE FOR PARSING
-multi_output_model.save('multi_output_model.h5')
+multi_output_model.save('multi_output_model.h5', save_format='h5')
 
 # CONVERT TO TXT
 convert_weights.h5_to_txt('multi_output_model.h5', 'single_output_model.txt')

KerasWeightsProcessing/examples/test_network.py

@@ -1,4 +1,4 @@
-import keras
+from tensorflow import keras
 import argparse
 import numpy as np
 import subprocess
@@ -12,11 +12,11 @@
 # set random seeds for reproducibility
 np.random.seed(123)
 import tensorflow as tf
-tf.set_random_seed(123)
+tf.compat.v1.set_random_seed(123)
 
-from keras.models import Sequential, Model
-from keras.layers import Dense, Input, LeakyReLU, Dropout, BatchNormalization
-from keras.models import load_model
+from tensorflow.keras.models import Sequential, Model
+from tensorflow.keras.layers import Dense, Input, LeakyReLU, Dropout, BatchNormalization
+from tensorflow.keras.models import load_model
 
 parser = argparse.ArgumentParser()
 parser.add_argument('--train', action='store_true')
@@ -111,7 +111,7 @@
 keras_predictions = model.predict(example_input)[0]
 
 # save the weights
-model.save(weights_file)
+model.save(weights_file, save_format='h5')
 # convert h5 file to txt
 h5_to_txt(
     weights_file_name=weights_file,