Skip to content

amita1996/Humpback-Whale-Identification

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

17 Commits
 
 
 
 

Repository files navigation

Humpback-Whale-Identification

As a part of the Data Science workshop at the Open University of Israel we tackled the problem of Humpback Whale Identification on Kaggle.

The dataset presents many problems such as many classes (5005 classes), few images for each class (41% of the classes with only 1 image), 38% of the dataset is unlabeled (labeled as "new whale" which is a whale we haven't seen before), different shapes and color of an image (RGB or Grayscale) and many more.

In our solution we used EfficientNetV2_S, ArcFace Loss and YoloV7 in order to achieve 88.5% accuracy on the test data.

Images of what our model learned: image

The notebook might be too big to display on github. Please use one of the following links: nbviewer link or Colab link

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published