Question

我尝试在tensorflow中定义一个二维占位符，但是，我事先并不知道它的大小。因此我定义了另一个占位符，但它似乎根本不起作用。这是最小的例子：

<script type="text/javascript">

$(document).ready(function(){

    $("#barScan").click(function(){
       cordova.plugins.barcodeScanner.scan(
     function (result) {
         alert("We got a barcode\n" +
               "Result: " + result.text + "\n" +
               "Format: " + result.format + "\n" +
               "Cancelled: " + result.cancelled);
     }, 
     function (error) {
         alert("Scanning failed: " + error);
     }
);
    });

});
 </script>

错误讯息：

import tensorflow as tf

batchSize = tf.placeholder(tf.int32)
input = tf.placeholder(tf.int32, [batchSize, 5])

然后我试着收拾形状，所以我有这个：

Traceback (most recent call last):
  File "C:/Users/v-zhaom/OneDrive/testconv/test_placeholder.py", line 5, in <module>
    input = tf.placeholder(tf.int32, [batchSize, 5])
  File "C:\Python35\lib\site-packages\tensorflow\python\ops\array_ops.py", line 1579, in placeholder
    shape = tensor_shape.as_shape(shape)
  File "C:\Python35\lib\site-packages\tensorflow\python\framework\tensor_shape.py", line 821, in as_shape
    return TensorShape(shape)
  File "C:\Python35\lib\site-packages\tensorflow\python\framework\tensor_shape.py", line 457, in __init__
    self._dims = [as_dimension(d) for d in dims_iter]
  File "C:\Python35\lib\site-packages\tensorflow\python\framework\tensor_shape.py", line 457, in <listcomp>
    self._dims = [as_dimension(d) for d in dims_iter]
  File "C:\Python35\lib\site-packages\tensorflow\python\framework\tensor_shape.py", line 378, in as_dimension
    return Dimension(value)
  File "C:\Python35\lib\site-packages\tensorflow\python\framework\tensor_shape.py", line 33, in __init__
    self._value = int(value)
TypeError: int() argument must be a string, a bytes-like object or a number, not 'Tensor'

也不起作用：

    input = tf.placeholder(tf.int32, tf.pack([batchSize, 5]))

Answer 1

如果您事先不知道某个维度的长度，请使用None，例如

input = tf.placeholder(tf.int32, [None, 5])

当你为这个占位符提供一个正确的形状数组（batch_size，5）时，它的动态形状将被正确设置，即

sess.run(tf.shape(input), feed_dict={input: np.zeros(dtype=np.int32, shape=(10, 5))})

将返回

array([10,  5], dtype=int32)

按预期

使用占位符作为张量流中的形状

1 个答案: