Tensorflow RNN输出张量形状

时间:2017-08-08 10:53:27

标签: python tensorflow neural-network shape recurrent-neural-network

我对张量流很陌生,我无法完全了解如何塑造张量,以便将输出作为单个数字。基本上,我的经常性网络应该猜测下一个数字。相反,对于每个预测,它返回一个包含五个数字的列表?我猜我的一个或多个张量都是畸形的。

我的输入数据被格式化为大约2000个列表,其中包含5个功能:

[
  np.array ([
              [1],[2],[3],[4],[5]
            ]) 
]

这是RNN的代码:

cell_units = 400
batch_size = 5
no_of_epochs = 500

data = tf.placeholder (tf.float32, [None, 5, 1])
target = tf.placeholder (tf.float32, [None, 1, 1])


weight = tf.Variable (tf.random_normal ([cell_units, 5, 1]))
bias = tf.Variable (tf.random_normal([1, 1]))


cell = tf.contrib.rnn.BasicRNNCell (num_units = cell_units)

output, states = tf.nn.dynamic_rnn (cell, data, dtype=tf.float32)


output = tf.transpose (output, [1, 0, 2])


activation = tf.matmul (output, weight) + bias


cost = tf.reduce_mean (
                        (
                            tf.log (tf.square (activation - target))
                        )
                      )

optimizer = tf.train.AdamOptimizer (learning_rate = 0.01).minimize(cost)

with tf.Session () as sess:

    sess.run (tf.global_variables_initializer ())
    no_of_batches = int (len (train_x) / batch_size)

    for i in range(no_of_epochs):
        start = 0
        for j in range(no_of_batches):
            inp = train_x [start:start+batch_size]
            out = train_y [start:start+batch_size]
            start += batch_size

            sess.run (optimizer, {data: inp, target: out})

1 个答案:

答案 0 :(得分:0)

tf.nn.dynamic_rnn期望输入形状[batch_size, max_time, ...]。在您的示例中,batch_size是动态的(即未知),max_time5(即时间步数)。当然,RNN的输出包含5条目,每个输入步骤一个:[None, 5, cell_units]

由于@ Ishant Mrinal建议您可以选择最后一个输出步骤。

weight = tf.Variable (tf.random_normal ([cell_units, 1]))
bias = tf.Variable (tf.random_normal([1, 1]))

cell = tf.contrib.rnn.BasicRNNCell (num_units = cell_units)
output, states = tf.nn.dynamic_rnn (cell, data, dtype=tf.float32)
# Get the last step (4th index).
output = tf.squeeze(tf.transpose (output, [0, 2, 1])[:,:,4]) # Shape of [batch_size, cell_units].
activation = tf.matmul (output, weight) + bias

activation的形状为[batch_size, 1]