search

Optimization-AI / LibAUC

LibAUC: A Deep Learning Library for X-Risk Optimization
https://libauc.org/
MIT License
285 stars 38 forks source link

Multi-Label Training #3

Closed Tom2096 closed 2 years ago

Tom2096 commented 2 years ago

Hello! I was wondering for the Multi-Label training example, do we need to use Sigmoid after getting the output from densenet 121? Also, is this Multi-Label training referring to the case where a single sample can have multiple classes, or is it the case where a single sample can only be one of many classes? Thanks!

yzhuoning commented 2 years ago

You do need to pass the outputs (logits) through a sigmoid layer before constructing loss. For multi-label, it means that one example (image) can have more than one label. In example (7), we focus on multi-label instead of multi-class.

Tom2096 commented 2 years ago

Thanks for the clarification!