search

SanghunYun / UDA_pytorch

UDA(Unsupervised Data Augmentation) implemented by pytorch
Apache License 2.0
275 stars 61 forks source link

Why unsup_loss is divided by unsup_loss_mask not total? #20

Open kgarg8 opened 2 years ago

kgarg8 commented 2 years ago

Hi SanghunYun,

Thanks for making the Pytorch version of UDA available.

Can I ask you the reason behind your dividing unsup_loss by number of ones in unsup_loss_mask and not total number of samples?

Reference: https://github.com/SanghunYun/UDA_pytorch/blob/0ba5cf8d8a6f698e19a295119f084a17dfa7a1e3/main.py#L129