TensorFlow中针对tf.GradientTape的急切执行错误

如何解决TensorFlow中针对tf.GradientTape的急切执行错误

当我尝试创建显着性可视化时收到此警告。

RuntimeError                              Traceback (most recent call last)

<ipython-input-36-6f13b9abef1d> in <module>()
      8   fig,axes = plt.subplots(1,2)
      9   # Generate visualization
---> 10   visualize = visualize_saliency(saved_model,layer_index,filter_indices=input_class,seed_input=input_image)
     11   axes[0].imshow(input_image[...,0])
     12   axes[0].set_title('Original image')


tf.gradients is not supported when eager execution is enabled. Use tf.GradientTape instead.

这是我的代码

layer_index = utils.find_layer_idx(saved_model,'dense_2')
saved_model.layers[layer_index].activation = activations.linear
saved_model = utils.apply_modifications(saved_model)  
indices_to_visualize = [ 0,12,38,83,112,74,190 ]

# Visualize
for index_to_visualize in indices_to_visualize:
  # Get input
  input_image = X_test[index_to_visualize]
  input_class = np.argmax(y_test[index_to_visualize])
  # Matplotlib preparations
  fig,2)
  # Generate visualization
  visualize = visualize_saliency(saved_model,seed_input=input_image)
  axes[0].imshow(input_image[...,0]) 
  axes[0].set_title('Original image')
  axes[1].imshow(visualize)
  axes[1].set_title('Saliency map')
  fig.suptitle(f'MNIST target = {input_class}')
  plt.show()

我尝试过在线查看其他类似的错误，但是我不明白如何在我的代码中实现这些解决方案。我应该如何在显着性可视化功能上使用tf.GradientTape。对于显着性可视化，我还没有遇到同样的错误。