Skip to content
Pytorch implementation of the paper "Class-Balanced Loss Based on Effective Number of Samples"
Python
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
LICENSE
README.md
class_balanced_loss.py
image.png
loss.png
samples.png

README.md

Class-balanced-loss-pytorch

Pytorch implementation of the paper Class-Balanced Loss Based on Effective Number of Samples presented at CVPR'19.

Yin Cui, Menglin Jia, Tsung-Yi Lin(Google Brain), Yang Song(Google), Serge Belongie

Dependencies

  • Python (>=3.6)
  • Pytorch (>=1.2.0)

Review article of the paper

Medium Article

How it works

It works on the principle of calculating effective number of samples for all classes which is defined as:

alt-text

Thus, the loss function is defined as:

alt-text

Visualisation for effective number of samples

alt-text

References

official tensorflow implementation

You can’t perform that action at this time.