Question

有一个程序定义了损失函数，如下所示：

def loss(hypes, decoded_logits, labels):
"""Calculate the loss from the logits and the labels.

Args:
  logits: Logits tensor, float - [batch_size, NUM_CLASSES].
  labels: Labels tensor, int32 - [batch_size].

Returns:
  loss: Loss tensor of type float.
"""
logits = decoded_logits['logits']
with tf.name_scope('loss'):
    logits = tf.reshape(logits, (-1, 2))
    shape = [logits.get_shape()[0], 2]
    epsilon = tf.constant(value=hypes['solver']['epsilon'])
    # logits = logits + epsilon
    labels = tf.to_float(tf.reshape(labels, (-1, 2)))

    softmax = tf.nn.softmax(logits) + epsilon

    if hypes['loss'] == 'xentropy':
        cross_entropy_mean = _compute_cross_entropy_mean(hypes, labels,
                                                         softmax)
    elif hypes['loss'] == 'softF1':
        cross_entropy_mean = _compute_f1(hypes, labels, softmax, epsilon)

    elif hypes['loss'] == 'softIU':
        cross_entropy_mean = _compute_soft_ui(hypes, labels, softmax,
                                              epsilon)



    reg_loss_col = tf.GraphKeys.REGULARIZATION_LOSSES

    print('******'*10)
    print('loss type ',hypes['loss'])
    print('type ', type(tf.get_collection(reg_loss_col)))
    print( "Regression loss collection: {}".format(tf.get_collection(reg_loss_col)))
    print('******'*10)


    weight_loss = tf.add_n(tf.get_collection(reg_loss_col))

    total_loss = cross_entropy_mean + weight_loss

    losses = {}
    losses['total_loss'] = total_loss
    losses['xentropy'] = cross_entropy_mean
    losses['weight_loss'] = weight_loss

return losses

运行程序会引发以下错误消息

File "/home/ decoder/kitti_multiloss.py", line 86, in loss
    name='reg_loss')
  File "/devl /tensorflow/tf_0.12/lib/python3.4/site-packages/tensorflow/python/ops/math_ops.py", line 1827, in add_n
    raise ValueError("inputs must be a list of at least one Tensor with the "
ValueError: inputs must be a list of at least one Tensor with the same dtype and shape

我检查了tf.add_n的功能，其实现如下。我的问题是如何检查tf.get_collection(reg_loss_col)中的第一个参数tf.add_n并打印其信息以找出错误消息的生成原因？

def add_n(inputs, name=None):
  """Adds all input tensors element-wise.
  Args:
    inputs: A list of `Tensor` objects, each with same shape and type.
    name: A name for the operation (optional).
  Returns:
    A `Tensor` of same shape and type as the elements of `inputs`.
  Raises:
    ValueError: If `inputs` don't all have same shape and dtype or the shape
    cannot be inferred.
  """
  if not inputs or not isinstance(inputs, (list, tuple)):
    raise ValueError("inputs must be a list of at least one Tensor with the "
                     "same dtype and shape")
  inputs = ops.convert_n_to_tensor_or_indexed_slices(inputs)
  if not all(isinstance(x, ops.Tensor) for x in inputs):
    raise ValueError("inputs must be a list of at least one Tensor with the "
                     "same dtype and shape")

Answer 1

为什么你甚至需要进入add_n才能看到tf.get_collection(reg_loss_col)是什么？您可以拥有tmp = tf.get_collection(reg_loss_col)，然后查看其类型。顺便说一句，看起来您的图表中没有任何正规化损失，在这种情况下tf.get_collection(reg_loss_col)将返回一个空列表。

在Python中查看对象的类型您可以使用内置函数type。例如，要查看tmp的类型：print type(tmp)

Answer 2

作为解决方法，您可以将以下行替换为：

temp = tf.get_collection('losses')

if temp == []:
            temp = [0]
        weight_loss = tf.add_n(temp, name='total_loss')

添加零值不会影响最终结果，但可以有效地运行软件...您如何看待？

关于ValueError：如果`inputs`不是都具有相同的形状和dtype或形状

2 个答案: