C4W3 Exercise 3 - unet_model

Course Q&A Deep Learning Specialization Convolutional Neural Networks
1

I don’t understand the concept of :
“chain the first output element of each block to the input of the next convolutional block”

Why it should be

cblock2 = conv_block(cblock1[0], n_filters * 2)
Instead of
cblock2 = conv_block(cblock1, n_filters * 2)

2

conv_block retuns (next_layer, skip_connection).

  1. next_layer refers to the output from the current conv block. It is the input to the next conv block and hence cblock1[0].
  2. The skip_connection will be used in the decoding path.

Hope this helps.