1 paper
Qi Li, Cheng-Long Wang, Yinzhi Cao +1
Training models on a carefully chosen portion of data rather than the full dataset is now a standard preprocess for modern ML. From vision coreset selection to large-scale filterin…