pith. machine review for the scientific record. sign in

arxiv: 1811.06992 · v2 · submitted 2018-11-16 · 💻 cs.LG · cs.DC· stat.ML

Recognition: unknown

Image Classification at Supercomputer Scale

Authors on Pith no claims yet
classification 💻 cs.LG cs.DCstat.ML
keywords optimizationsaccuracybatchdeeplearningscalethroughputtraining
0
0 comments X
read the original abstract

Deep learning is extremely computationally intensive, and hardware vendors have responded by building faster accelerators in large clusters. Training deep learning models at petaFLOPS scale requires overcoming both algorithmic and systems software challenges. In this paper, we discuss three systems-related optimizations: (1) distributed batch normalization to control per-replica batch sizes, (2) input pipeline optimizations to sustain model throughput, and (3) 2-D torus all-reduce to speed up gradient summation. We combine these optimizations to train ResNet-50 on ImageNet to 76.3% accuracy in 2.2 minutes on a 1024-chip TPU v3 Pod with a training throughput of over 1.05 million images/second and no accuracy drop.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.