Hi, I'm interested in this challenge and I have a question for limits of strong augmentation models like GAN models.
According to the Q&A section of ROBIN-dataset github, we can only use provided training dataset and ImageNet-1K dataset.
But whether we can use the additional dataset for strong data augmentations (e.g. GANs) are not mentioned exactly.
So, I'm curious about we can use the additional dataset to train strong data augmentation model like GANs.
Posted by: wooks527 @ Aug. 7, 2022, 1:37 a.m.Hi,
Thanks for your interest in our challenge.
And the answer is no additional data is allowed for data augmentation, if you choose to augment the data in anyway, please only use the provided training set and ImageNet-1k dataset
Thank you for your prompt reply.
Posted by: wooks527 @ Aug. 7, 2022, 10:41 a.m.