Export Citations
2 Results for: Author: Pauloski, J Gregory
Save this search
Please login to be able to save your searches and receive alerts for new content matching your search criteria.
Searched The ACM Full-Text Collection (691,749 records)|Expand your search to The ACM Guide to Computing Literature (3,482,418 records)
Showing 1 - 2of2 Results
Select All
- research-articleNovember 2021
KAISA: an adaptive second-order optimizer framework for deep neural networks
SC '21: Proceedings of the International Conference for High Performance Computing, Networking, Storage and AnalysisNovember 2021, Article No.: 13, pp 1–14https://doi.org/10.1145/3458817.3476152Kronecker-factored Approximate Curvature (K-FAC) has recently been shown to converge faster in deep neural network (DNN) training than stochastic gradient descent (SGD); however, K-FAC's larger memory footprint hinders its applicability to large models. ...
- research-articleNovember 2020
Convolutional neural network training with distributed K-FAC
SC '20: Proceedings of the International Conference for High Performance Computing, Networking, Storage and AnalysisNovember 2020, Article No.: 94, pp 1–14Training neural networks with many processors can reduce time-to-solution; however, it is challenging to maintain convergence and efficiency at large scales. The Kronecker-factored Approximate Curvature (K-FAC) was recently proposed as an approximation ...
