Why chain the elements rather than just input the whole tensor? [DLS4 Week 3 Unet assignment]

In the downsampling and upsampling cancatenate, why use the first element as input?
If only chain the elements, will all the elements in the tensor get processed?
Why not input the whole tensor rather than chaining them?
Look forward to any help. Thank you.

Posting solution code in a public topic is discouraged and can get your account suspended. It’s okay to share stacktrace on a public post and send code to a mentor via direct message. Please clean up the post.

Chaining refers to the logical flow of data. These are actually actual function calls.
For custom functions, the body within the function definition is invoked. For tensorflow related calls, __call__ gets invoked on the layer instance.

As far as concatenating input tensors is concerned, do see the operations inside functions like upsampling_block does this operation internally. Depending on how the model layers are arranged, you’ll feed the data accordingly.

Thanks for your reminder and explaination.