search

SwinTransformer / Transformer-SSL

This is an official implementation for "Self-Supervised Learning with Swin Transformers".
https://arxiv.org/abs/2105.04553
MIT License
629 stars 67 forks source link

The interpolation method for BYOL augmentation is wrong #12

Open songkangsg opened 2 years ago

songkangsg commented 2 years ago

Under Transformer-SSL/data/build.py, inside the "build_transform" function, under "byol" augmentation type, the interpolation method used in RandomResizedCrop is the default which is BILINEAR, however in the BYOL paper the author used BICUBIC

Giles-Billenness commented 2 years ago

Yeah I saw this as well, BILINEAR is less effective as well, although hopefully, it wouldn't have much of an effect.