Hi! Thank you for this awesome paper. My team and I have been using SwAV heavily in our work. I am trying to train your model from scratch on ImageNet and was wondering if you use the ILSVRC 1.2M subset, or if you use the entire dataset (~14M images). Many thanks!
Hi! Thank you for this awesome paper. My team and I have been using SwAV heavily in our work. I am trying to train your model from scratch on ImageNet and was wondering if you use the ILSVRC 1.2M subset, or if you use the entire dataset (~14M images). Many thanks!