The official implementation of [CVPR2022] Decoupled Knowledge Distillation https://arxiv.org/abs/2203.08679 and [ICCV2023] DOT: A Distillation-Oriented Trainer https://openaccess.thecvf.com/content/ICCV2023/papers/Zhao_DOT_A_Distillation-Oriented_Trainer_ICCV_2023_paper.pdf
Hello, I want to know how the acc in paper calculated?
In paper, I found All results are the average over 5/3 trials on CIFAR100/ImageNet, and I want to know whether the acc is the best acc or something else.
Many thanks!
Our paper reports the best accuracies among epochs. As for large-scale datasets such as MS-COCO and ImageNet, the best accuracy is pretty much the same as the accuracy of the last epoch.
Hello, I want to know how the acc in paper calculated? In paper, I found
All results are the average over 5/3 trials on CIFAR100/ImageNet
, and I want to know whether the acc is the best acc or something else. Many thanks!