Fig. 10
From: Deep representation learning using layer-wise VICReg losses

Performance of the model trained on EMNIST as training continues.
From: Deep representation learning using layer-wise VICReg losses

Performance of the model trained on EMNIST as training continues.