search

YyzHarry / imbalanced-semi-self

[NeurIPS 2020] Semi-Supervision (Unlabeled Data) & Self-Supervision Improve Class-Imbalanced / Long-Tailed Learning
https://arxiv.org/abs/2006.07529
MIT License
736 stars 116 forks source link

command to get a base classifier in semi-supervised learning #20

Closed abababa-ai closed 11 months ago

abababa-ai commented 3 years ago

In semi-supervised training, what is the command to get the base classifier using only labeled data?

YyzHarry commented 3 years ago

Hi - thanks for your interest. The base model is obtained by normal training procedure on the labeled imbalanced dataset. You can also customize the base loss function or training procedure for the base classifier. For example, with CE, \rho=100 on CIFAR-10:

python train.py --dataset cifar10 --imb_factor 0.01 --loss_type CE

Or with LDAM:

python train.py --dataset cifar10 --imb_factor 0.01 --loss_type LDAM