python - TensorFlow dynamic_rnn 回归量 : ValueError dimension mismatch

Question

我想构建一个用于回归的玩具 LSTM 模型。 This不错的教程对于初学者来说已经太复杂了。

给定一个长度为 time_steps 的序列，预测下一个值。考虑 time_steps=3 和序列:

array([
   [[  1.],
    [  2.],
    [  3.]],

   [[  2.],
    [  3.],
    [  4.]],
    ...

目标值应该是:

array([  4.,   5., ...

我定义了以下模型:

# Network Parameters
time_steps = 3 
num_neurons= 64 #(arbitrary)
n_features = 1

# tf Graph input
x = tf.placeholder("float", [None, time_steps, n_features])
y = tf.placeholder("float", [None, 1])

# Define weights
weights = {
   'out': tf.Variable(tf.random_normal([n_hidden, 1]))
} 
biases = {
   'out': tf.Variable(tf.random_normal([1]))
}

#LSTM model
def lstm_model(X, weights, biases, learning_rate=0.01, optimizer='Adagrad'):

  # Prepare data shape to match `rnn` function requirements
  # Current data input shape: (batch_size, time_steps, n_features)
  # Required shape: 'time_steps' tensors list of shape (batch_size, n_features)
  # Permuting batch_size and time_steps
  input dimension: Tensor("Placeholder_:0", shape=(?, 3, 1), dtype=float32)

  X = tf.transpose(X, [1, 0, 2])
  transposed dimension: Tensor("transpose_41:0", shape=(3, ?, 1), dtype=float32)

  # Reshaping to (time_steps*batch_size, n_features)
  X = tf.reshape(X, [-1, n_features])
  reshaped dimension: Tensor("Reshape_:0", shape=(?, 1), dtype=float32)

  # Split to get a list of 'time_steps' tensors of shape (batch_size, n_features)
  X = tf.split(0, time_steps, X)
  splitted dimension: [<tf.Tensor 'split_:0' shape=(?, 1) dtype=float32>, <tf.Tensor 'split_:1' shape=(?, 1) dtype=float32>, <tf.Tensor 'split_:2' shape=(?, 1) dtype=float32>]

  # LSTM cell
  cell = tf.nn.rnn_cell.LSTMCell(num_neurons) #Or GRUCell(num_neurons)

  output, state = tf.nn.dynamic_rnn(cell=cell, inputs=X, dtype=tf.float32)

  output = tf.transpose(output, [1, 0, 2])
  last = tf.gather(output, int(output.get_shape()[0]) - 1)


  return tf.matmul(last, weights['out']) + biases['out']

我们用 pred = lstm_model(x, weights, biases) 实例化 LSTM 模型我得到以下信息:

---> output, state = tf.nn.dynamic_rnn(cell=cell, inputs=X, dtype=tf.float32)
ValueError: Dimension must be 2 but is 3 for 'transpose_42' (op: 'Transpose') with input shapes: [?,1], [3]

1)你知道问题出在哪里吗？

2) 将 LSTM 输出乘以权重会产生回归吗？

Answer 1

正如评论中所讨论的，tf.nn.dynamic_rnn(cell, inputs, ...)函数需要一个三维张量列表^* 作为其inputs 参数，其中维度默认解释为batch_size x num_timesteps x num_features。 (如果您传递 time_major=True，它们将被解释为 num_timesteps x batch_size x num_features。)因此预处理你在原始占位符中所做的是不必要的，你可以将 oriding X 值直接传递给 tf.nn.dynamic_rnn()。

---

^* 技术上它可以接受除列表之外的复杂嵌套结构，但叶子元素必须是三维张量。^**

^** 对此进行调查后发现了 tf.nn.dynamic_rnn() 实现中的错误。原则上，输入至少有两个维度就足够了，但是 time_major=False 路径在将输入转置为时间主要形式时假定它们恰好具有三个维度，并且这是这个错误无意中导致出现在您的程序中的错误消息。我们正在努力解决这个问题。