Skip to content

NickDGreg/pk-batching-for-triplet-loss-tutorial

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

6 Commits
 
 
 
 
 
 
 
 

Repository files navigation

pk-batching-for-triplet-loss-tutorial

The concept of triplet loss training is best described in this blog post by Oliver Moindrot and this tensorflow tutorial trains a model using it.

This tutorial is based on In Defense of the Triplet Loss for Person Re-Identification, a paper which shows how to avoid some of the difficulties of triplet loss training. In it they create batches of P classes and K examples from each class to maximise the training speed.

What I show here is how to create Tensorflow Records with the classes and samples arranged in PK batches so you can optimise your triplet loss training.

About

A jupyter notebook showing how to create PK batches for optimal triplet loss training

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published