10/8/2022
•
EN
Not Every Model Has a Separate "Loss Function"
Argues that separating 'loss functions' from models is a flawed abstraction in deep learning systems, proposing a more flexible design.