Running the Tensorflow 2.0 code gives 'ValueError: tf.function-decorated function tried to create variables on non-first call'. What am I doing wrong?

Question

可以看出，我在 'error_giving_notebook' 中使用了 tf.function 装饰器，它抛出一个 ValueError 而同一个笔记本除了删除 tf.function 装饰器外没有任何变化运行在 'non_problematic_notebook' 中顺利进行。可能是什么原因？

Answer 1

这里的问题出在class conv2d的调用方法的return值中：

if self.bias:
  if self.pad == 'REFLECT':
    self.p = (self.filter_size - 1) // 2
    self.x = tf.pad(inputs, [[0, 0], [self.p, self.p], [self.p, self.p], [0, 0]], 'REFLECT')
    return Conv2D(filters=self.filter_num, kernel_size=(self.filter_size, self.filter_size), strides=(self.stride, self.stride),
                                  padding='VALID', use_bias=True, kernel_initializer=self.w, bias_initializer=self.b)(self.x)
  else:
    return Conv2D(filters=self.filter_num, kernel_size=(self.filter_size, self.filter_size), strides=(self.stride, self.stride),
                                  padding=self.pad, use_bias=True, kernel_initializer=self.w, bias_initializer=self.b)(inputs)
else:
   if self.pad == 'REFLECT':
      self.p = (self.filter_size - 1) // 2
      self.x = tf.pad(inputs, [[0, 0], [self.p, self.p], [self.p, self.p], [0, 0]], 'REFLECT')
      return Conv2D(filters=self.filter_num, kernel_size=(self.filter_size, self.filter_size), strides=(self.stride, self.stride),
                                  padding='VALID', use_bias=False, kernel_initializer=self.w)(self.x)
   else:
      return Conv2D(filters=self.filter_num, kernel_size=(self.filter_size, self.filter_size), strides=(self.stride, self.stride),
                                  padding=self.pad, use_bias=False, kernel_initializer=self.w)(inputs)

通过 return 创建一个 Conv2D 对象 tf.Variable(s) 每次你调用

predictions = model(images)

在你的 tf 修饰函数中。因此，异常。

解决此问题的一种可能方法是更改 conv2d class 中的构建和调用方法，如下所示：

def build(self, inputs):
  self.w = tf.random_normal_initializer(mean=0.0, stddev=1e-4)
  if self.bias:
    self.b = tf.constant_initializer(0.0)
  else:
    self.b = None

  self.conv_a = Conv2D(filters=self.filter_num, kernel_size=(self.filter_size, self.filter_size), strides=(self.stride, self.stride), padding='VALID', use_bias=True, kernel_initializer=self.w, bias_initializer=self.b)
  self.conv_b = Conv2D(filters=self.filter_num, kernel_size=(self.filter_size, self.filter_size), strides=(self.stride, self.stride), padding=self.pad, use_bias=True, kernel_initializer=self.w, bias_initializer=self.b)
  self.conv_c = Conv2D(filters=self.filter_num, kernel_size=(self.filter_size, self.filter_size), strides=(self.stride, self.stride), padding='VALID', use_bias=False, kernel_initializer=self.w)
  self.conv_d = Conv2D(filters=self.filter_num, kernel_size=(self.filter_size, self.filter_size), strides=(self.stride, self.stride),padding=self.pad, use_bias=False, kernel_initializer=self.w)  

def call(self, inputs):
  if self.bias:
    if self.pad == 'REFLECT':
      self.p = (self.filter_size - 1) // 2
      self.x = tf.pad(inputs, [[0, 0], [self.p, self.p], [self.p, self.p], [0, 0]], 'REFLECT')
      return self.conv_a(self.x)
    else:
      return self.conv_b(inputs)
  else:
     if self.pad == 'REFLECT':
        self.p = (self.filter_size - 1) // 2
        self.x = tf.pad(inputs, [[0, 0], [self.p, self.p], [self.p, self.p], [0, 0]], 'REFLECT')
        return self.conv_c(self.x)
     else:
        return self.conv_d(inputs)

为了更好地理解 AutoGraph 以及@tf.function 的工作原理，我建议您查看 this

Answer 2

由于您尝试在 TF 2.0 中使用函数装饰器，请在导入 TensorFlow 后使用以下行热切启用运行函数：

tf.config.experimental_run_functions_eagerly(True)

由于以上内容已弃用（不再是实验性的？），请改用以下内容：

tf.config.run_functions_eagerly(True)

如果您想了解更多信息，请参阅 this link。

Running the Tensorflow 2.0 code gives 'ValueError: tf.function-decorated function tried to create variables on non-first call'. What am I doing wrong?

Running the Tensorflow 2.0 code gives 'ValueError: tf.function-decorated function tried to create variables on non-first call'. What am I doing wrong?

deep-learning

keras

keras-layer

tf.keras

tensorflow2.0