q512624756/aifood_modelarts

aifood recognition

• large dataset：75 class

• small dataset：25 class

method

• MobileNetV3
• center loss + softmax loss
• Use Huawei ModelArts platform training model

Reference

• https://github.com/KaiyangZhou/pytorch-center-loss

  Wen et al. A Discriminative Feature Learning Approach for Deep Face Recognition. ECCV 2016.

• https://github.com/xiaolai-sqlai/mobilenetv3

  paper: https://arxiv.org/pdf/1905.02244.pdf