i have a convolutional neural network that has 3 conovlutional layers, followed by one dense layer with 64 nodes.
When I change the dropout to 0.99, that should make the netwrok to crash, my accuracy is still above 90%.
Why is that ?
# flatten output and feed into dense layer
model.add(tf.keras.layers.Flatten())
model.add(tf.keras.layers.Dense(64, activation='relu'))
tf.keras.layers.Dropout(0.3)
Thank you in advance.