Assignment - Tensor Flow - Using One Hot Encoding

I have a suggestion that the class labels are listed under each columns to reflect the elements in the row vector of y.

The current class = 0, class =1, class = 2, class = 3, makes a confusion for me to take each row match a label.

Please ignore this if it is only an issue for me.


Hi TimGTian,

I can see where the confusion is coming from, and I think is valid. However, that’s isn’t exacty what the picture is trying to convey.

In the picture, and as you noted, each row is a class. It is just that is left to the student to figure out that the position of the ‘1’ is what marks each column as a member of the class.

In other words, imagine you only had 1 column. You would only have one ‘1’ and it would be positioned in the row index that matched it’s class.

Hope that makes sense.

I’ll raise this to the maintainers to see if we could add a little note with an explanation below that picture.