English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

With an increasing demand for training powers for deep learning algorithmsand the rapid growth of computation resources in data centers, it is desirableto dynamically schedule different distributed deep learning tasks to maximizeresource utilization and reduce cost. In this process, different tasks mayreceive varying numbers of machines at different time, a setting we callelastic distributed training. Despite the recent successes in large mini-batchdistributed training, these methods are rarely tested in elastic distributedtraining environments and suffer degraded performance in our experiments, whenwe adjust the learning rate linearly immediately with respect to the batchsize. One difficulty we observe is that the noise in the stochastic momentumestimation is accumulated over time and will have delayed effects when thebatch size changes. We therefore propose to smoothly adjust the learning rateover time to alleviate the influence of the noisy momentum estimation. Ourexperiments on image classification, object detection and semantic segmentationhave demonstrated that our proposed Dynamic SGD method achieves stabilizedperformance when varying the number of GPUs from 8 to 128. We also providetheoretical understanding on the optimality of linear learning rate schedulingand the effects of stochastic momentum.

Dynamic Mini-batch SGD for Elastic Distributed Training: Learning in the Limbo of Resources