1 code implementation • 30 Sep 2018 • Sangkug Lym, Armand Behroozi, Wei Wen, Ge Li, Yongkee Kwon, Mattan Erez
Training convolutional neural networks (CNNs) requires intense computations and high memory bandwidth.
no code implementations • 2 Apr 2019 • Sangkug Lym, Donghyuk Lee, Mike O'Connor, Niladrish Chatterjee, Mattan Erez
Training convolutional neural networks (CNNs) requires intense compute throughput and high memory bandwidth.
no code implementations • 6 Mar 2019 • Esha Choukse, Michael Sullivan, Mike O'Connor, Mattan Erez, Jeff Pool, David Nellans, Steve Keckler
However, GPU device memory tends to be relatively small and the memory capacity can not be increased by the user.
Hardware Architecture
no code implementations • 27 Apr 2020 • Sangkug Lym, Mattan Erez
Based on our evaluation, FlexSA with the proposed compilation heuristic improves compute resource utilization of pruning and training modern CNN models by 37% compared to a conventional training accelerator with a large systolic array.
no code implementations • 10 Jun 2020 • Benjamin Ghaemmaghami, Zihao Deng, Benjamin Cho, Leo Orshansky, Ashish Kumar Singh, Mattan Erez, Michael Orshansky
Increasing the dimension of embedding vectors improves model accuracy but comes at a high cost to model size.
no code implementations • 27 Sep 2023 • Zihao Deng, Benjamin Ghaemmaghami, Ashish Kumar Singh, Benjamin Cho, Leo Orshansky, Mattan Erez, Michael Orshansky
At constant model quality, MLET allows embedding dimension, and model size, reduction by up to 16x, and 5. 8x on average, across the models.
no code implementations • 2 Oct 2023 • Yeonsoo Jeon, Mattan Erez, Michael Orshansky
Privacy-Preserving ML (PPML) based on Homomorphic Encryption (HE) is a promising foundational privacy technology.
1 code implementation • 26 Jan 2019 • Sangkug Lym, Esha Choukse, Siavash Zangeneh, Wei Wen, Sujay Sanghavi, Mattan Erez
State-of-the-art convolutional neural networks (CNNs) used in vision applications have large models with numerous weights.