Keras error: expected dense_input_1 to have 3 dimensions

The problem lies here:

model.add(Dense(32, input_shape=(5, 3)))

it should be:

model.add(Dense(32, input_shape=(3,)))

This first example dimension is not included in input_shape. Also because it's actually dependent on batch_size set during network fitting. If you want to specify you could try:

model.add(Dense(32, batch_input_shape=(5, 3)))

EDIT:

From your comment I understood that you want your input to have shape=(5,3) in this case you need to:

1. reshape your x by setting:

   x = x.reshape((1, 5, 3))
   

   where first dimension comes from examples.

2. You need to flatten your model at some stage. It's because without it you'll be passing a 2d input through your network. I advice you to do the following:

   model = Sequential()
   model.add(Dense(32, input_shape=(5, 3)))
   model.add(Activation('relu'))
   model.add(Dense(32))
   model.add(Activation('relu'))
   model.add(Flatten())
   model.add(Dense(4))