Ungraded Lab 1 Step 2d: Why transpose before concat?

Hi @Rajiv_Bharadwaja

The Step 2d is needed for the else statement to be equivalent with 'tf,fhb->htb' (check this post).

While the Step 3 is just concatenating the rotated_vecs with the same but opposite sign values (see the same linked post).

In other words, the 2a, 2b, 2c and 2d steps are the implementation of the 'tf,fhb->htb' matrix multiplication and Step 3 concatenation on the “irrelevant” axis is logical and in no way in odds with the previous steps.

I hope that makes sense :slight_smile: