Skip to content

Cross domain few-shot transfer learning from MiniImageNet to EuroSAT_RGB and CUB

License

Notifications You must be signed in to change notification settings

KRVPerera/FSL-MiniImageNet

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

27 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

CDFSL MiniImageNet to EuroSAT and CUB

Pretraining notebooks

Finetuning notebooks

EuroSAT_RGB

CUB

Accuracy of models

Pre-train on miniImageNet dataset

Model Validation Accuracy
VGG19 82.84%
VisionTransformer 93.83%

Fine tuned on EuroSAT_RGB dataset

  • Episodes : 20
Model Average Test Accuracy
VGG19 78.94%
VisionTransformer 81.34%

Fine tuned on CUB dataset

  • Episodes : 20
Model Average Test Accuracy
VGG19 77.067%
VisionTransformer 81.34%

Fine tuned on CUB dataset

  • Episodes : 50
Model Average Test Accuracy
VGG19 73.41%
VisionTransformer 93.34%

Pretrained - models

RestNet152d - best_model_ModelResnet152dTimm_FixedTransforms.pth VGG19 - best_model_VGG19_fixedTransformers.pth VisionTransformer - best_model_VistionTransformerTimm_pc18_fixedTransformers.pth

You can download pretrained models from this location https://drive.google.com/drive/folders/1oD5kTRVJPSjhVz2C6cJeI1EXc2tmHMWq

About

Cross domain few-shot transfer learning from MiniImageNet to EuroSAT_RGB and CUB

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages