Skip to content

sararahman1729/Custom-Loss-Function

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 

Repository files navigation

Custom-Loss-Function

This project introduces a custom loss function for Knowledge Distillation (KD) technique to improve the training of the student model.
In this project a custom loss function using the standard cross entropy loss function and confidence penalty is being created to test in some datasets like CIFAR-100, Tiny ImageNet, CUB-200.

Student Models Used: ShuffleNetV1, ResNet-18, ResNet-34, ResNet-50, EfficientNet-B0.

Teacher Models Used: VGG13-BN, ResNet50, DeepLabV3+ResNet101 (for segmentation), EfficientNet-B0/B3, ResNet-50.

About

I did this project during my Neural network and pattern recognition course.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors