Skip to content

Latest commit

 

History

History
16 lines (16 loc) · 755 Bytes

README.md

File metadata and controls

16 lines (16 loc) · 755 Bytes

MF-MGVC

  • Code release for 《Multi-Frequency Feature Enhancement for Multi-Granularity Visual Classification》

Requirements

  • python 3.6
  • PyTorch 1.2.0
  • torchvision

Data

  • Download datasets
  • Extract them to data/Birds/, data/Airs/ and data/Cars/, respectively.
  • Split the dataset into train and test folder, the index of each class should follow the /MGVC dataset list

Training

  • Just run main.py
  • The parameters you see are training CUB-200-2011 dataset in main.py. If you want to run Stanford Cars and FGVC-Aircraft datasets, you can adjust them directly in the main.py

Contact

Thanks for your attention! If you have any suggestion or question, you can leave a message here or contact us directly: