CNN Using TensorFlow

Convolutional Neural Network/CNNtf.py

@@ -0,0 +1,30 @@
+
+import tensorflow as tf
+import matplotlib.pyplot as plt
+#plt.imshow(x_train[0])
+plt.imshow(x_train[0],cmap=plt.cm.binary)
+plt.show()
+print(x_test[0])
+
+mnist = tf.keras.datasets.mnist  # 28x28 images of handwritten digits from 0-9
+(x_train,y_train),(x_test,y_test) = mnist.load_data()  # distibution labels of input data
+
+x_train = tf.keras.utils.normalize(x_train,axis=1) # normalization of training data
+x_test = tf.keras.utils.normalize(x_test,axis=1)  # normalization of testing data
+
+model = tf.keras.models.Sequential()  
+model.add(tf.keras.layers.Flatten())  
+model.add(tf.keras.layers.Dense(128,activation=tf.nn.relu)) 
+model.add(tf.keras.layers.Dense(128,activation=tf.nn.relu))
+model.add(tf.keras.layers.Dense(10,activation=tf.nn.softmax)) 
+
+
+model.compile (optimizer='adam',
+               loss="sparse_categorical_crossentropy",
+               metrics =['accuracy']
+               )
+
+model.fit(x_train,y_train,epochs=3)
+
+val_loss, val_accuracy = model.evaluate(x_test,y_test)
+print(val_loss,val_accuracy)