Video: Coding Contrastive Loss

Course Q&A TensorFlow: Advanced Techniques Specialization Custom Models, Layers and Loss Functions with TF
1

Hi,
I have a question about the coding of the Class statement which appears around 1:29 in the video of “Coding Contrastive Loss”.
Why do we have the following line before initializing attributes?

margin = 0
For your reference, I attach herewith the screenshot of the Video.

W2 Contrastive Loss Class
W2 Contrastive Loss Class1920×1200 245 KB

I do not see any point in fixing the value of margin within the Class statement, especially when we want to parameterize margin.
How does it affect when we assign another value to the parameter, margin, in the function?
I would appreciate it if anyone can advise me on this issue.
Thanks in advance.
Best regards
Michio

2

Having the line margin = 0 at the top level of the class defines it as a class variable, but, I agree, there’s no particular reason I can think of to have a class variable for margin when there’s also an instance variable for it. One possibility is that the line just happens to be left-over from an earlier version the instructor was working with before making margin a parameter to init.

3

Thanks @Wendy for the clarification!
So I can skip it when I write the code!
Thanks
Best regards
Michio