Timothée Darcet

> > @Suhail To generate features from the pretrained backbones, just use a transform similar to the standard one used for evaluating on image classification with the typical ImageNet normalization...

> The fact that the layer is linear does not really matter, it is just a way to say that DINOv2's frozen features are really good, so that you can...

> In the paper it says that iBoT loss is used for distillation. If I understand correctly, the distillation's iBoT is done with predicting a distribution with the teacher given...

Hi, My advice would be to not load the weights using the checkpointer resume function, as it also expects the optimizer buffer etc. However you can launch a normal training...