Cropr uses removable auxiliary heads to learn task-relevant token pruning in ViTs, achieving 1.5-4x speedups with small accuracy drops across image classification, semantic segmentation, and object detection.
Soft-nms–improving object detection with one line of code
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
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Token Cropr: Faster ViTs for Quite a Few Tasks
Cropr uses removable auxiliary heads to learn task-relevant token pruning in ViTs, achieving 1.5-4x speedups with small accuracy drops across image classification, semantic segmentation, and object detection.