Pith. sign in

Cross-image relational knowledge distillation for semantic segmentation,

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it

citation-role summary

background 1

citation-polarity summary

fields

cs.CV 1

years

2025 1

verdicts

CONDITIONAL 1

roles

background 1

polarities

unclear 1

representative citing papers

DeepKD: A Deeply Decoupled and Denoised Knowledge Distillation Trainer

cs.CV · 2025-05-21 · conditional · novelty 6.0

DeepKD is a knowledge distillation trainer that decouples task, target-class, and non-target-class gradients with GSNR-based momentum and a dynamic top-k mask, yielding consistent accuracy gains on CIFAR-100, ImageNet, and MS-COCO.

citing papers explorer

Showing 1 of 1 citing paper.

  • DeepKD: A Deeply Decoupled and Denoised Knowledge Distillation Trainer cs.CV · 2025-05-21 · conditional · none · ref 3

    DeepKD is a knowledge distillation trainer that decouples task, target-class, and non-target-class gradients with GSNR-based momentum and a dynamic top-k mask, yielding consistent accuracy gains on CIFAR-100, ImageNet, and MS-COCO.