tf.keras.layers.LSTM(8, input_shape=x_train_uni.shape[-2:])

tf.keras.layers.LSTM(8, input_shape=x_train_uni.shape[-2:]) 最后面的shape [-2:] 代表什么意思呢？