in Keras 2.2.0 model have different layer index

[jason9075]

In Keras 2.1.0
vgg.summary()

_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
input_1 (InputLayer)         (None, 397, 635, 3)       0         
_________________________________________________________________
block1_conv1 (Conv2D)        (None, 397, 635, 64)      1792      
_________________________________________________________________
block1_conv2 (Conv2D)        (None, 397, 635, 64)      36928     
_________________________________________________________________
........
........
........

But in Keras 2.2.0

vgg.summary()

_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
block1_conv1 (Conv2D)        (None, 397, 635, 64)      1792      
_________________________________________________________________
block1_conv2 (Conv2D)        (None, 397, 635, 64)      36928     
_________________________________________________________________
average_pooling2d_1 (Average (None, 198, 317, 64)      0         
_________________________________________________________________
........
........
........

We find that "input_1 (InputLayer) " was missing. So the index of layers were also changed.