Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Data processing method, data processing device and electronic device

A data processing device and data processing technology, applied in the computer field, can solve the problem of low acceleration, achieve the effect of increasing the acceleration ratio, shortening the waiting time, and shortening the time consumed

Active Publication Date: 2018-06-29
TENCENT TECH (SHENZHEN) CO LTD
View PDF7 Cites 12 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] However, the applicant found that: currently, multiple GPUs are used for parallel processing for deep learning training, and its acceleration is relatively low

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
  • Data processing method, data processing device and electronic device
  • Data processing method, data processing device and electronic device
  • Data processing method, data processing device and electronic device

Examples

Experimental program
Comparison scheme
Effect test

Embodiment approach

[0130] As an implementation manner, the correction parameter transmission control unit 301 includes a first parameter exchange control unit, and the model parameter distribution control unit 50 includes a first distribution control unit.

[0131] in:

[0132] The first parameter exchange control unit is used to control each slave GPU to transmit the correction parameters generated during training and not exchanged to the master GPU through the peripheral interface.

[0133] The first distribution control unit is configured to control the master GPU to transmit the corrected model parameters to the slave GPUs respectively through the peripheral interface.

[0134] Optionally, the first parameter exchange control unit is specifically configured to: control each slave GPU to convert the non-exchanged correction generated by training to The parameters are transferred to the main GPU through the peripheral interface.

[0135] Optionally, the training process control unit 20 is spec...

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 data processing method. The method includes loading different training data set to a plurality of GPUs; controlling the GPUs to run in parallel so as to the received trainingdata sets, and storing modification parameters generated in a reveries process to corresponding display memories; in a process of controlling the GPUs to run in parallel so as to train the received training data set, controlling the GPUs to perform exchange treatment on the modification parameters generated through training and not subjected to exchanging. According to the invention, when the GPUs implement training of the training data sets, a part of modification parameters generated in training of each GPU is exchanged to other GPUs, so that waiting time of the GPUs can be shortened, timeconsumed in each round can be shorted, and further time consumed in the whole depth learning training can be shortened. The equipment acceleration ratio is improved. The invention also discloses a corresponding data processing device and an electronic device.

Description

technical field [0001] The invention belongs to the technical field of computers, and in particular relates to a data processing method, a data processing device and electronic equipment. Background technique [0002] Deep learning originated from the research of artificial neural networks, and the multi-layer perceptron with multiple hidden layers is a deep learning structure. Deep learning combines low-level features to form more abstract high-level representation attribute categories or features to achieve distributed feature representation of data. For example, the convolutional neural network architecture (Convolutional Architecture for FastFeature Embedding, referred to as caffe) is widely used in image recognition and text recognition. [0003] GPU (Graphics Processing Unit), also known as graphics processor, display core or display chip, is a microprocessor that performs image calculations on personal computers, workstations, game consoles and some mobile devices (s...

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): G06N99/00
CPCG06N20/00G06N20/20G06N5/01G06N7/01G06T1/20G06T1/60G06N99/00G06V10/955G06F18/214
Inventor 高进朝李毅薛伟金涬
Owner TENCENT TECH (SHENZHEN) CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com