TensorFlow 仅在使用 MultiRNNCell 时抛出错误

Question

我正在使用传统的序列到序列框架在 TensorFlow 1.0.1 中构建编码器-解码器模型。当我在编码器和解码器中有一层 LSTM 时，一切正常。但是，当我尝试将 >1 层的 LSTM 包裹在 MultiRNNCell 中时，调用 tf.contrib.legacy_seq2seq.rnn_decoder 时出现错误。

完整的错误在这篇文章的最后，但简而言之，它是由一行引起的

(c_prev, m_prev) = state

在 TensorFlow 中抛出 TypeError: 'Tensor' object is not iterable.。我对此感到困惑，因为我传递给rnn_decoder 的初始状态确实是一个元组。据我所知，使用 1 层或 >1 层的唯一区别是后者涉及使用MultiRNNCell。使用它时我应该知道一些 API 怪癖吗？

这是我的代码（基于this GitHub repo 中的示例）。为它的长度道歉；这是我可以做到的最小限度，同时仍然是完整和可验证的。

import tensorflow as tf
import tensorflow.contrib.legacy_seq2seq as seq2seq
import tensorflow.contrib.rnn as rnn

seq_len = 50
input_dim = 300
output_dim = 12
num_layers = 2
hidden_units = 100

sess = tf.Session()

encoder_inputs = []
decoder_inputs = []

for i in range(seq_len):
    encoder_inputs.append(tf.placeholder(tf.float32, shape=(None, input_dim),
                                         name="encoder_{0}".format(i)))

for i in range(seq_len + 1):
    decoder_inputs.append(tf.placeholder(tf.float32, shape=(None, output_dim),
                                         name="decoder_{0}".format(i)))

if num_layers > 1:
    # Encoder cells (bidirectional)
    # Forward
    enc_cells_fw = [rnn.LSTMCell(hidden_units)
                    for _ in range(num_layers)]
    enc_cell_fw = rnn.MultiRNNCell(enc_cells_fw)
    # Backward
    enc_cells_bw = [rnn.LSTMCell(hidden_units)
                    for _ in range(num_layers)]
    enc_cell_bw = rnn.MultiRNNCell(enc_cells_bw)
    # Decoder cell
    dec_cells = [rnn.LSTMCell(2*hidden_units)
                 for _ in range(num_layers)]
    dec_cell = rnn.MultiRNNCell(dec_cells)
else:
    # Encoder
    enc_cell_fw = rnn.LSTMCell(hidden_units)
    enc_cell_bw = rnn.LSTMCell(hidden_units)
    # Decoder
    dec_cell = rnn.LSTMCell(2*hidden_units)

# Make sure input and output are the correct dimensions
enc_cell_fw = rnn.InputProjectionWrapper(enc_cell_fw, input_dim)
enc_cell_bw = rnn.InputProjectionWrapper(enc_cell_bw, input_dim)
dec_cell = rnn.OutputProjectionWrapper(dec_cell, output_dim)

_, final_fw_state, final_bw_state = \
     rnn.static_bidirectional_rnn(enc_cell_fw,
                                  enc_cell_bw,
                                  encoder_inputs,
                                  dtype=tf.float32)

# Concatenate forward and backward cell states
# (The state is a tuple of previous output and cell state)
if num_layers == 1:
    initial_dec_state = tuple([tf.concat([final_fw_state[i],
                                          final_bw_state[i]], 1) 
                               for i in range(2)])
else:
    initial_dec_state = tuple([tf.concat([final_fw_state[-1][i],
                                          final_bw_state[-1][i]], 1) 
                               for i in range(2)])

decoder = seq2seq.rnn_decoder(decoder_inputs, initial_dec_state, dec_cell)

tf.global_variables_initializer().run(session=sess)

这是错误：

Traceback (most recent call last):
  File "example.py", line 67, in <module>
    decoder = seq2seq.rnn_decoder(decoder_inputs, initial_dec_state, dec_cell)
  File "/home/tao/.virtualenvs/example/lib/python2.7/site-packages/tensorflow/contrib/legacy_seq2seq/python/ops/seq2seq.py", line 150, in rnn_decoder
    output, state = cell(inp, state)
  File "/home/tao/.virtualenvs/example/lib/python2.7/site-packages/tensorflow/contrib/rnn/python/ops/core_rnn_cell_impl.py", line 426, in __call__
    output, res_state = self._cell(inputs, state)
  File "/home/tao/.virtualenvs/example/lib/python2.7/site-packages/tensorflow/contrib/rnn/python/ops/core_rnn_cell_impl.py", line 655, in __call__
    cur_inp, new_state = cell(cur_inp, cur_state)
  File "/home/tao/.virtualenvs/example/lib/python2.7/site-packages/tensorflow/contrib/rnn/python/ops/core_rnn_cell_impl.py", line 321, in __call__
    (c_prev, m_prev) = state
  File "/home/tao/.virtualenvs/example/lib/python2.7/site-packages/tensorflow/python/framework/ops.py", line 502, in __iter__
    raise TypeError("'Tensor' object is not iterable.")
TypeError: 'Tensor' object is not iterable.

谢谢！

Answer 1

问题在于传递给seq2seq.rnn_decoder的初始状态（initial_dec_state）的格式。

当您使用rnn.MultiRNNCell 时，您正在构建一个多层循环网络，因此您需要为这些层中的每个提供一个初始状态。

因此，您应该提供一个元组的列表作为初始状态，其中列表的每个元素都是来自循环网络相应层的先前状态。

所以你的initial_dec_state，初始化如下：

    initial_dec_state = tuple([tf.concat([final_fw_state[-1][i],
                                      final_bw_state[-1][i]], 1) 
                           for i in range(2)])

应该是这样的：

    initial_dec_state = [
                    tuple([tf.concat([final_fw_state[j][i],final_bw_state[j][i]], 1) 
                           for i in range(2)]) for j in range(len(final_fw_state))
                        ]

创建一个元组列表，格式如下：

    [(state_c1, state_m1), (state_c2, state_m2) ...]

更详细地说，'Tensor' object is not iterable. 错误的发生是因为 seq2seq.rnn_decoder 在内部调用了您的 rnn.MultiRNNCell (dec_cell)，并将初始状态 (initial_dec_state) 传递给它。

rnn.MultiRNNCell.__call__ 遍历初始状态列表，并为每个状态提取元组 (c_prev, m_prev)（在语句 (c_prev, m_prev) = state 中）。

所以如果你只传递一个元组，rnn.MultiRNNCell.__call__ 将遍历它，一旦它到达(c_prev, m_prev) = state，它就会找到一个张量（应该是一个元组）state 并抛出 @ 987654339@错误。

了解seq2seq.rnn_decoder 期望的初始状态格式的好方法是调用dec_cell.zero_state(batch_size, dtype=tf.float32)。此方法以初始化您正在使用的循环模块所需的确切格式返回零填充状态张量。