<ul data-eligibleForWebStory="true"><li>AdaDeDup is a hybrid data pruning framework designed to enhance the efficiency of training large-scale object detection models by integrating density-based pruning with model-informed feedback.</li><li>The framework partitions data, applies initial density-based pruning, and uses a proxy model to adjust cluster-specific pruning thresholds adaptively based on the impact of pruning on losses within each cluster.</li><li>Extensive experiments on Waymo, COCO, and nuScenes datasets using standard models show that AdaDeDup outperforms existing methods, reduces performance degradation, and maintains model performance while pruning 20% of data.</li><li>AdaDeDup's effectiveness in improving data efficiency for large-scale model training is highlighted by achieving near-original model performance with significant data reduction.</li></ul>

AdaDeDup: Adaptive Hybrid Data Pruning for Efficient Large-Scale Object Detection Training

Discover more