meaning of "Depth" argument

[I-CANT-CODE]

I just have a clarification question, when I specify a depth of 3 for the neural network, from observing the code I think that means that this network will have 3 down convolutions and 3 up convolutions, for a total of 5 U-Net blocks, rather than 3 UNET blocks total. Is this correct thinking?

[clberube]

I realize this question is over a year old but it may help others who are learning.

Please have a look at the original U-Net architecture below.
If depth = 5, then the original U-Net is reproduced. The code loops over depthand makes 5 down-blocks. Then it loops over depth-1 to make 4 up-blocks, for a total of 9 blocks.
If you use depth = 3, I believe you are correct thinking that the result will be 3 down-blocks + 2 up-blocks, for a total of 5.

Cheers.