Views
No views yet
mrl : ResNet50 mrl models trained with Matryoshka loss (vanilla and efficient) with nesting starting from d=8 (default) and d=16fixed-feature : independently trained ResNet50 baselines at log(d) granularitiesresnet-family : mrl and ff models trained on ResNet18/34/101@inproceedings{kusupati2022matryoshka,
title = {Matryoshka Representation Learning},
author = {Kusupati, Aditya and Bhatt, Gantavya and Rege, Aniket and Wallingford, Matthew and Sinha, Aditya and Ramanujan, Vivek and Howard-Snyder, William and Chen, Kaifeng and Kakade, Sham and Jain, Prateek and others},
title = {Matryoshka Representation Learning.},
booktitle = {Advances in Neural Information Processing Systems},
month = {December},
year = {2022},
}