Recent years, many applications have been driven advances by the use of Machine Learning (ML). Nowadays, it is common to see industrial-strength machine learning jobs that involve millions of model parameters, terabytes of training data, and weeks of training. Good efficiency, i.e., fast completion time of running a specific ML job, therefore, is a key feature of a successful ML system. While the completion time of a longrunning ML job is determined by the time required to reach model convergence, practically that is also largely influenced by the values of various system settings. In this paper, we contribute techniques towards building self-tuning parameter servers. Parameter Server (PS) is a popular system architecture for large-scale machine learning systems; and by self-tuning we mean while a long-running ML job is iteratively training the expert-suggested model, the system is also iteratively learning which system setting is more efficient for that job and applies it online. While our techniques are general enough to various PSstyle ML systems, we have prototyped our techniques on top of TensorFlow. Experiments show that our techniques can reduce the completion times of a variety of long-running TensorFlow jobs from 1.4× to 18×.
Comply with the development trend of the domestic mobile Internet and mobile devices and boom, in order to improve and optimize the school teachers and students working and life in Guilin University of Technology, to promote the exchange of teachers and students, to enrich the lives of teachers and students, especially research and design of Cloud-based campus-plus system on android platform, the Android client school Education Online system perfect butt, perfect integration with the Kingsoft cloud network disk, docking with the microblogging platform based on the Hadoop technology of cloud and some personalization features. Through the Firebug plug-in for Firefox browser Senate online system packet capture analysis, to get access to the data request URL and request parameters, and then to use HttpClient open source library to encapsulate the URL and parameters, initiates a connection request to the server, the last Jsoup and Gson open source library to parse and filter the data to develop a Senate online system client. Hadoop cloud microblogging server using the HTTPS protocol as well as the Json format for communication to transfer data, and also incorporates Kingsoft cloud network disk, in order to meet the daily lives of teachers and students anytime, anywhere communication, share resources with each other.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.