Ragged tensor as input in LSTM for word classification

Question

I builded a lstm model for one to one word classification:

model = Sequential()
model.add(layers.Input(shape=(None, 512), ragged=True))
model.add(layers.LSTM(32,