我是一个有状态的RNN与嵌入层,以执行二进制分类。我对函数API中需要的batch_size和batch_shape有一些混淆。h0 object has the same shape as RNN layer hidden states obtained when return_state = True.(xte_pad) >> (2400, 1403, 1) (2
使用TensorFlow (1.13.1)中的多层/堆叠RNN的初始状态所需的结构是什么?我尝试了以下几点:lstm_cells = [tf.keras.layers.LSTMCell(size) for size instate_init = [tf.placeholder(tf.float32, shape=[None] + cell.state_size) for cell in ls