A public cloud-based distributed deep learning task resource and batch size collaborative configuration method

By establishing a training time and loss value prediction model and combining it with the tabu search algorithm to optimize resource and batch size configuration, the training cost problem of distributed deep learning tasks in a public cloud environment is solved, and efficient training that reaches the predetermined loss value within a limited time is achieved.

CN115952420BActive Publication Date: 2026-06-09BEIJING UNIV OF TECH
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
BEIJING UNIV OF TECH
Filing Date
2023-01-06
Publication Date
2026-06-09

AI Technical Summary

Technical Problem

In a public cloud environment, the training cost of distributed deep learning tasks is high and it is difficult to reach the predetermined loss value within a limited time. Existing optimization schemes fail to effectively coordinate the allocation of resources and batch size, affecting training efficiency and cost.

Method used

Establish a model to predict training time and training loss value, combine it with the tabu search algorithm to optimize resource and batch size configuration, and find a co-configuration solution that minimizes training cost through mathematical modeling and machine learning techniques.

Benefits of technology

It effectively reduces the training cost of distributed deep model training tasks while ensuring training results, optimizes resource and batch size configuration, and improves training efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115952420B_ABST
    Figure CN115952420B_ABST
Patent Text Reader

Abstract

The application designs a resource and batch size collaborative configuration method for distributed deep model training tasks based on public cloud. First, the training process of different training modes is analyzed, and the training time is expressed by using parameters such as resources and batch size to establish a training time prediction model. Secondly, considering the influence of resources and batch size on the training loss value and the characteristic that the loss value curve can be approximated as an inverse proportional function, the method uses an inverse proportional function to approximate the loss value convergence curve, and uses resources, batch size and other features that affect the training loss value convergence as input to predict the parameters in the inverse proportional function using a machine learning model to establish a training loss value prediction model. Finally, the method uses the two prediction models above as the constraint conditions for the search to find the resource and batch size collaborative configuration solution that can minimize the training cost. The application can effectively reduce the training cost on the basis of reaching the specified loss value within a limited time.
Need to check novelty before this filing date? Find Prior Art