Faster Neural Network Training with Data Echoing
Dami Choi, Alexandre Passos, Christopher J. Shallue et al. · arXiv (Cornell University) · 2019
In the twilight of Moore's law, GPUs and other specialized hardware accelerators have dramatically sped up neural network training. However, earlier stages of the training pipeline, such as disk I/O and data preprocessing, do not run on accelerators. As accelerators continue to improve, these earlier stages will increasingly become the bottleneck. In this paper, we introduce "data echoing," which reduces the total computation used by earlier pipeline stages and speeds up training whenever computation upstream from accelerators dominates the training time. Data echoing reuses (or "echoes") intermediate outputs from earlier pipeline stages in order to reclaim idle capacity. We investigate the behavior of different data echoing algorithms on various workloads, for various amounts of echoing, and for various batch sizes. We find that in all settings, at least one data echoing algorithm can match the baseline's predictive performance using less upstream computation. We measured a factor of 3.25 decrease in wall-clock time for ResNet-50 on ImageNet when reading training data over a network.
Cited by 29 works. doi:10.48550/arxiv.1907.05550 Open-access PDF
Authors
References
- Deep Residual Learning for Image Recognition
- Learning representations by back-propagating errors
- ImageNet Large Scale Visual Recognition Challenge
- Learning Multiple Layers of Features from Tiny Images
- Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
- Human-level control through deep reinforcement learning
- SSD: Single Shot MultiBox Detector
- Attention Is All You Need
- On the importance of initialization and momentum in deep learning
- Parallelized Stochastic Gradient Descent