Unlock instant, AI-driven research and patent intelligence for your innovation.

Distributed framework applied to AI algorithm engineering and systematization

An algorithm engineering and distributed technology, applied in special data processing applications, computing, resource allocation, etc., can solve the problems of time-consuming hardware resources, inability to achieve high availability, and inability to expand hardware capabilities indefinitely, to reduce configuration requirements. Effect

Inactive Publication Date: 2017-08-11
SHANGHAI JILIAN NETWORK TECH CO LTD
View PDF5 Cites 7 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Generally speaking, using deep learning to analyze and recognize videos is an offline task that consumes a lot of time and hardware resources; the hardware capabilities of a single machine cannot be infinitely expanded to meet large batches of video analysis, nor can it achieve high availability

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Distributed framework applied to AI algorithm engineering and systematization
  • Distributed framework applied to AI algorithm engineering and systematization

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0017] The entire distributed framework is divided into four parts: task queue and scheduling service, video cutting service, algorithm analysis service, data center, and docker packaging for the convenience of distributed deployment of algorithms. The role of each service is described below, and finally the flow of the entire task can be clearly seen through the system framework diagram and flow chart.

[0018] 1. Task queue and scheduling system.

[0019] After the video file content is input, it will first enter the task queue and wait for the scheduler to process it. The implementation of the queue system can use rabbitmq or other tools that provide queue function implementation. The main purpose of using the queue method is to ensure that all other back-end services are not impacted when encountering a large number of video inputs; at the same time, the current task volume and status can be clearly observed through the background of the scheduling system to help operatio...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a distributed framework applied to AI algorithm engineering and systematization. The whole distributed framework is composed of four major parts: task queuing and scheduling services, video segmentation services, algorithm analysis services and a data center. The distributed framework can reduce the configuration requirements of a single machine, can effectively utilize the computation capability of each machine through the scheduling system, can accelerate analysis of a single video, and can perform lateral extension response to analysis tasks of a large batch of videos.

Description

technical field [0001] The present invention provides a distributed framework for accelerating deep learning video analysis and effectively improving the utilization rate of computer cpu and gpu, and is specifically applied to a distributed framework for engineering and systematization of AI algorithms. Background technique [0002] The distributed framework applied to AI algorithm engineering and systemization provided by the present invention is used to accelerate deep learning video analysis and effectively improve the utilization rate of computer cpu and gpu. Generally speaking, using deep learning to analyze and recognize videos is an offline task that consumes a lot of time and hardware resources; the hardware capabilities of a single machine cannot be infinitely expanded to meet large-scale video analysis, nor can it achieve high availability. In response to such a situation, a distributed architecture must be adopted to expand the capabilities of the machine, increas...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06N99/00G06F9/50G06F17/30
CPCG06F9/5027G06F16/70G06N20/00
Inventor 刘聪
Owner SHANGHAI JILIAN NETWORK TECH CO LTD