Codes of Nexperia image classification project
Including using CNN models (Resnet50) with cost-sensitive learning (LDAM, CB-focal and BSCE loss); Vision transformers, including DeiT, CaiT, ResT and Swin-Transformer; Self-supervised learning pretraining (MoCo)