June 10, 2019
Distributed data-parallel algorithms aim to accelerate the training of deep neural networks by parallelizing the computation of large mini-batch gradient updates across multiple nodes. Approaches that synchronize nodes using exact distributed averaging (e.g., via ALLREDUCE) are sensitive to stragglers and communication delays. The PUSHSUM gossip algorithm is robust to these issues, but only performs approximate distributed averaging. This paper studies Stochastic Gradient Push (SGP), which combines PUSHSUM with stochastic gradient updates. We prove that SGP converges to a stationary point of smooth, non-convex objectives at the same sub-linear rate as SGD, and that all nodes achieve consensus. We empirically validate the performance of SGP on image classification (ResNet-50, ImageNet) and machine translation (Transformer, WMT’16 EnDe) workloads. Our code will be made publicly available.
Research Topics
November 10, 2022
Unnat Jain, Abhinav Gupta, Himangi Mittal, Pedro Morgado
November 10, 2022
November 06, 2022
Filip Radenovic, Abhimanyu Dubey, Dhruv Mahajan
November 06, 2022
October 25, 2022
Mustafa Mukadam, Austin Wang, Brandon Amos, Daniel DeTone, Jing Dong, Joe Ortiz, Luis Pineda, Maurizio Monge, Ricky Chen, Shobha Venkataraman, Stuart Anderson, Taosha Fan, Paloma Sodhi
October 25, 2022
October 22, 2022
Naila Murray, Lei Wang, Piotr Koniusz, Shan Zhang
October 22, 2022
April 30, 2018
Yedid Hoshen, Lior Wolf
April 30, 2018
December 11, 2019
Eliya Nachmani, Lior Wolf
December 11, 2019
April 30, 2018
Yedid Hoshen, Lior Wolf
April 30, 2018
November 01, 2018
Yedid Hoshen, Lior Wolf
November 01, 2018
Foundational models
Our approach
Latest news
Foundational models