A memory allocation method and system for effectively accelerating deep learning calculations
A technology of memory allocation and deep learning, applied in the field of deep learning system optimization, can solve problems such as complexity, achieve the effect of reducing the size of occupied memory space and improving computing efficiency
Active Publication Date: 2021-04-16
VASTAI TECH SHANGHAI INC
View PDF8 Cites 1 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
With the deepening of deep learning research, the number of layers of artificial neural networks is increasing, from 8 layers of AlexNet to 19 layers of VGG and 22 layers of GoogLeNet, and even ResNet is as deep as 152 layers. Gradually more multi-branch networks appear, wider and more complex
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 moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0017] In order to enable those skilled in the art to better understand the technical solutions of the present application, the following will clearly and completely describe the technical solutions in the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. Obviously, the described implementation Examples are only some of the embodiments of the present application, not all of them. Based on the embodiments in this application, all other embodiments obtained by persons of ordinary skill in the art without creative efforts shall fall within the scope of protection of this application.
[0018] It should be noted that the terms "first" and "second" in the description and claims of the present application and the above drawings are used to distinguish similar objects, but not necessarily used to describe a specific sequence or sequence. It is to be understood that the data so used are interchangeable under appropri...
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
Login to View More Abstract
The present application provides a memory allocation method and system for effectively accelerating deep learning. The method and system can determine the target operation sequence of the multi-branch operation layer according to the size of the memory space required for the operation of the multi-branch operation layer; The target operation sequence executes the target memory allocation plan for the continuous storage of each branch operation result of the multi-branch operation layer; according to the target memory allocation plan and the target operation sequence, determine the memory allocation plan for executing the multi-branch operation layer . Therefore, when performing multi-branch operations, the operation sequence that occupies the smallest memory space during the operation process can be selected as the target operation sequence, and the multi-branch operations are carried out layer by layer according to the target memory scheme in which the results of each branch of the multi-branch operation layer are continuously stored. The memory allocation setting ensures that the calculation results of each branch are continuously stored in the memory space, thereby reducing the size of the occupied memory space and improving the calculation efficiency of the entire neural network.
Description
technical field [0001] This application relates to the field of deep learning system optimization, in particular to a memory allocation method and system for effectively accelerating deep learning calculations based on multi-branch scheduling and allocation. Background technique [0002] In recent years, artificial intelligence research has been very popular, and deep learning is one of the core technologies, and its basic model is deep neural network. With the deepening of deep learning research, the number of layers of artificial neural networks is increasing, from 8 layers of AlexNet to 19 layers of VGG and 22 layers of GoogLeNet, and even ResNet is as deep as 152 layers. Gradually more multi-branch networks emerged, wider and more complex. Deeper and wider networks mean that more memory is required to train the network model. [0003] Since the overall deeper and wider network has a significant effect on the accuracy of deep learning, the deep neural network is develop...
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
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G06F9/50G06N3/08
CPCG06F9/5016G06N3/08
Inventor 李国亮张磊杨勤富钱军
Owner VASTAI TECH SHANGHAI INC