Experimental KERAS 3

Signed-off-by: Chaurasiya, Payal <[email protected]>

openfl-workspace/keras_cnn_mnist/requirements.txt

@@ -1 +1,2 @@
-tensorflow==2.13
+keras==3.6
+tensorflow==2.18

openfl-workspace/keras_cnn_mnist/src/keras_cnn.py

@@ -3,11 +3,9 @@
 
 """You may copy this file as the starting point of your own model."""
 
-import tensorflow.keras as ke
-from tensorflow.keras import Sequential
-from tensorflow.keras.layers import Conv2D
-from tensorflow.keras.layers import Dense
-from tensorflow.keras.layers import Flatten
+import keras as ke
+from keras import Sequential
+from keras.layers import Conv2D, Dense, Flatten
 
 from openfl.federated import KerasTaskRunner
 
@@ -72,14 +70,14 @@ def build_model(self,
 
         model.add(Dense(num_classes, activation='softmax'))
 
-        model.compile(loss=ke.losses.categorical_crossentropy,
-                      optimizer=ke.optimizers.legacy.Adam(),
+        model.compile(loss=ke.losses.BinaryCrossentropy,
+                      optimizer=ke.optimizers.Adam(),
                       metrics=['accuracy'])
 
-        # initialize the optimizer variables
-        opt_vars = model.optimizer.variables()
-
-        for v in opt_vars:
-            v.initializer.run(session=self.sess)
+        # # initialize the optimizer variables
+        # opt_vars = model.optimizer.variables()
 
+        # for v in opt_vars:
+        #     v.initializer.run(session=self.sess)
+        #     # ke.backend.get_session().run(v.initializer)
         return model

openfl-workspace/keras_cnn_mnist/src/mnist_utils.py

@@ -4,9 +4,8 @@
 """You may copy this file as the starting point of your own model."""
 
 from logging import getLogger
-
 import numpy as np
-from tensorflow.python.keras.utils.data_utils import get_file
+import keras
 
 logger = getLogger(__name__)
 
@@ -38,18 +37,7 @@ def _load_raw_datashards(shard_num, collaborator_count):
     Returns:
         2 tuples: (image, label) of the training, validation dataset
     """
-    origin_folder = 'https://storage.googleapis.com/tensorflow/tf-keras-datasets/'
-    path = get_file('mnist.npz',
-                    origin=origin_folder + 'mnist.npz',
-                    file_hash='731c5ac602752760c8e48fbffcf8c3b850d9dc2a2aedcf2cc48468fc17b673d1')
-
-    with np.load(path) as f:
-        # get all of mnist
-        X_train_tot = f['x_train']
-        y_train_tot = f['y_train']
-
-        X_valid_tot = f['x_test']
-        y_valid_tot = f['y_test']
+    (X_train_tot, y_train_tot), (X_valid_tot, y_valid_tot )  = keras.datasets.mnist.load_data(path="mnist.npz")
 
     # create the shards
     shard_num = int(shard_num)
@@ -116,3 +104,5 @@ def load_mnist_shard(shard_num, collaborator_count, categorical=True,
         y_valid = one_hot(y_valid, num_classes)
 
     return input_shape, num_classes, X_train, y_train, X_valid, y_valid
+
+# _, num_classes, X_train, y_train, X_valid, y_valid = load_mnist_shard(0,2)