AkiraTOSEI / ML_papers

ML_paper_summary(in Japanese)
4 stars 1 forks source link

Unsupervised Learning of Visual Features by Contrasting Cluster Assignments #74

Open AkiraTOSEI opened 4 years ago

AkiraTOSEI commented 4 years ago

TL;DR

Whereas SimCLR and others require large batch sizes in representation learning, SwAV makes it possible to learn in small batches by online training that train the model to cluster consistently between differently augmented two image . Good results in object detection and image classification using the obtained representations. image

Why it matters:

Paper URL

https://arxiv.org/abs/2006.09882

Submission Dates(yyyy/mm/dd)

Authors and institutions

Methods

Results

Comments