Skip to content

Comparing different decision tree algorithms for the FashionMINST classification problem

Notifications You must be signed in to change notification settings

danielegana/FashionMNISTtreesandtransformers

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 

Repository files navigation

FashionMNIST with decision trees and transformers

Comparing different decision tree algorithms for the FashionMNIST classification problem

I compare:

  • Regular decision trees, accounting for seed dependency
  • Random forests
  • Scikit's bagging algorithm
  • XGboost
  • SimpleVIT Transformer!

Contrary to popular belief, trees do give a pretty good image classification accuracy that rivals CNNs, up to 90%, and with essentially no data preprocessing or normalization.

SimpleViT is slow, and I didn't manage to get it past 90% accuracy.

About

Comparing different decision tree algorithms for the FashionMINST classification problem

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages