tensorflow错误记录:DeprecationWarning: elementwise

在用tensorflow学习简单的神经网络时，报如下错误：

DeprecationWarning: elementwise == comparison failed; this will raise an error

原因是：在处理数据时，我运行了如下代码后，又重新粘贴了一份，再次运行，使得数据变成了三维，造成了数据的不匹配：

代码：

image_size = 28
num_labels = 10

def reformat(dataset, labels):
  dataset = dataset.reshape((-1, image_size * image_size)).astype(np.float32)
  # Map 0 to [1.0, 0.0, 0.0 ...], 1 to [0.0, 1.0, 0.0 ...]
  labels = (np.arange(num_labels) == labels[:,None]).astype(np.float32)
  return dataset, labels
train_dataset, train_labels = reformat(train_dataset, train_labels)
valid_dataset, valid_labels = reformat(valid_dataset, valid_labels)
test_dataset, test_labels = reformat(test_dataset, test_labels)
print('Training set', train_dataset.shape, train_labels.shape)
print('Validation set', valid_dataset.shape, valid_labels.shape)
print('Test set', test_dataset.shape, test_labels.shape)

正确结果：

Training set (200000, 784) (200000, 10)
Validation set (10000, 784) (10000, 10)
Test set (10000, 784) (10000, 10)

运行错误结果：

Training set (200000, 784) (200000, 1, 10)
Validation set (10000, 784) (10000, 1, 10)
Test set (10000, 784) (10000, 1, 10)

很明显，数据的维度有错

解决方法：重新运行第一遍的程序即可，只允许一遍即可。