• A
  • A
  • A
  • ABC
  • ABC
  • ABC
  • А
  • А
  • А
  • А
  • А
Regular version of the site

Cloud-based Service for Training Deep Neural Networks

Student: Krutoy Nikita

Supervisor: Oleg V. Sukhoroslov

Faculty: Faculty of Computer Science

Educational Programme: Applied Mathematics and Information Science (Bachelor)

Year of Graduation: 2019

Deep Neural Networks are popular instrument for modeling different types of systems. They consist of large number of parameters and require considerable amount of data to train on. Training time is measured in hours or days. Parallel methods are often used to speed up neural network training on multiprocessor systems. At the same time, there are distributed system, like grid systems, which have potential of achieving greater parallelism by using unlimited amount of workers. This work proposes a library with three methods for training deep neural networks on grid-systems. This work also includes the full description of the proposed library, set of experiments on training two models from two computer vision tasks and comparison of training speed-up achieved by using libraries method to a single processor system.

Student Theses at HSE must be completed in accordance with the University Rules and regulations specified by each educational programme.

Summaries of all theses must be published and made freely available on the HSE website.

The full text of a thesis can be published in open access on the HSE website only if the authoring student (copyright holder) agrees, or, if the thesis was written by a team of students, if all the co-authors (copyright holders) agree. After a thesis is published on the HSE website, it obtains the status of an online publication.

Student theses are objects of copyright and their use is subject to limitations in accordance with the Russian Federation’s law on intellectual property.

In the event that a thesis is quoted or otherwise used, reference to the author’s name and the source of quotation is required.

Search all student theses