jeroenoudevrielink / spark Goto Github PK
View Code? Open in Web Editor NEWThis project forked from keyu-tian/spark
[ICLR'23 Spotlight🔥] The first successful BERT/MAE-style pretraining on any convolutional network; Pytorch impl. of "Designing BERT for Convolutional Networks: Sparse and Hierarchical Masked Modeling"
Home Page: https://arxiv.org/abs/2301.03580
License: MIT License