Wasserstein-distance-based logit and feature distillation losses outperform KL-divergence-based counterparts on image classification and object detection.
Function-consistent feature distillation
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 1years
2024 1verdicts
ACCEPT 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Wasserstein Distance Rivals Kullback-Leibler Divergence for Knowledge Distillation
Wasserstein-distance-based logit and feature distillation losses outperform KL-divergence-based counterparts on image classification and object detection.