Dataset Distillation via Knowledge Distillation: Towards Efficient Self-Supervised Pre-training of Deep Networks

Official repository for "Dataset Distillation via Knowledge Distillation: Towards Efficient Self-Supervised Pre-training of Deep Networks" which presents the method MKDT for dataset distillation for self-supervised learning.

Obtaining Teacher Model

Generating Expert Trajectories

python 

Distilling Dataset

python 

Evaluating Distilled Dataset

python 

Bibtex

n/a