Refactor: Extract set_output Function from utils.h - #4467
Conversation
|
Rather than randomly fixing the tech-debts, it'd be better if we can create a issue and track what needs to be done there. We can create sub-tickets under that for a tech debt that needs attention |
|
Thanks for the recommendations. Usually I would create an issue and fix it however sometime I just randomly explore like this time lolll. So I don't really have a list of "tech-debt" task lolll |
a4109bf to
70c0a1b
Compare
zcbenz
left a comment
There was a problem hiding this comment.
While this does reduce duplicate code, it does not really make code more readable, specifically with the new relax_prev_stride arg it would take more time for a reader to understand what the code exactly does. Different kernels can have very slight different requirements on inputs and it is hard to extract common functions without harming readability.
For this case, it would make sense moving set_ouput to a common function within the normalization.cpp file though.
|
gonna do it tn thx for the comment. |
This PR extracts set_output function from normalization.cpp and softmax.cpp.
In total there are 3 same function call but using lambda function currently. This refactor allows us to reuse and reduce the complexity from the code.
Umm I used AI to move the the block to utils.h and understand this function especially the following line haha and also use some it to help me do the formatting and running the testing (I'm lazy to type the command loll)
no_copy &= (s == 0 || s == x.shape().back() || x.shape(-2) == 1);