Peeling the Onion: Hierarchical Reduction of Data Redundancy for Efficient Vision Transformer Training.
Zhenglun KongHaoyu MaGeng YuanMengshu SunYanyue XiePeiyan DongXin MengXuan ShenHao TangMinghai QinTianlong ChenXiaolong MaXiaohui XieZhangyang WangYanzhi WangPublished in: CoRR (2022)
Keyphrases
- data sets
- data collection
- redundant data
- complex data
- database
- data sources
- data quality
- data processing
- image data
- data reduction
- original data
- probability distribution
- data analysis
- real time
- databases
- fuzzy logic
- historical data
- computer vision
- training examples
- end users
- test data
- data distribution
- neural network
- machine learning
- social networks
- raw data
- training data
- supervised learning
- knowledge discovery
- training set