WO2024230170A1 - Quantization precision selection method and device for deep network model - Google Patents

Abstract

The present application relates to the technical field of deep learning, and in particular, to a quantization precision selection method and device for a deep network model. The quantization precision selection method comprises: S110, acquiring a deep network model to be processed, and letting the quantization precision of each layer in the deep network model be first precision; S120, determining the quantization sensitivity of each layer in the deep network model; S130, traversing, in the deep network model, layers the quantization precision of which is not changed, so as to select the layer having the lowest quantization sensitivity, and changing the quantization precision of the layer to be second precision, wherein the second precision is smaller than the first precision; and S140, performing inference by using the deep network model the quantization precision of which has been changed, and if the inference time is shorter than or equal to a preset target time, ending to obtain an expected deep network model. According to the present embodiment, the model quantization efficiency is improved while the precision of a network model is ensured.

Description

A method and device for selecting quantization accuracy of deep network model

This application claims priority to a Chinese patent application filed with the China Patent Office on May 11, 2023, with application number 2023105342377, and invention name “A method and device for selecting quantization accuracy of a deep network model”, the entire contents of which are incorporated by reference in this application.

Technical Field

The present application relates to the field of deep learning technology, and in particular to a method and device for selecting quantization accuracy of a deep network model.

Background Art

In the existing deep network deployment process, model quantization is an indispensable step. Common sense says that the higher the quantization bit width, the better the quantization effect, and the lower the quantization bit width, the worse the quantization accuracy. However, high bit width means longer inference time, and low bit width often results in poor accuracy of the model after quantization. Naturally, people began to study mixed-precision quantization tools. Most of the current mixed-precision quantization uses reinforcement learning methods to automatically determine the quantization accuracy of each layer.

However, existing tools with mixed-precision quantization functions pay too little attention to mixed precision. These tools often focus on the selection of quantization parameters and then determine the quantization bit width; or the quantization bit width is selected together with the quantization parameter, which obviously does not pay much attention to the selection of precision. At the same time, methods based on deep learning also face the problems of long quantization time and complex tool structure.

Summary of the invention

In view of this, an embodiment of the present application provides a method and device for selecting quantization accuracy of a deep network model, which can solve at least one technical problem in the related art.

In the first aspect, an embodiment of the present application provides a method for selecting quantization accuracy of a deep network model, including: S110, obtaining an input image and a deep network model to be processed, quantizing the parameters of the deep network model and setting the quantization accuracy of each layer in the quantized deep network model to a first accuracy; S120, inputting the input image into the quantized deep network model to obtain a feature map output by each layer of the deep network model, and determining the quantization sensitivity of each layer in the deep network model using the feature map output by each layer of the deep network model; S130, traversing the layers in the deep network model whose quantization accuracy has not been changed to select a layer with the smallest quantization sensitivity, and changing the quantization accuracy of the layer to a second accuracy, which is less than the first accuracy; S140, performing reasoning using the deep network model after the quantization accuracy has been changed, and if the reasoning time is greater than the preset target time, returning to step S130; if the reasoning time is less than or equal to the preset target time, then ending to obtain a deep network model that meets expectations. This embodiment adjusts the quantization accuracy of the network model in combination with the quantization sensitivity of each layer of the deep network model, while ensuring the accuracy of the network model, it improves the model quantization efficiency.

In some embodiments, determining the quantization sensitivity of each layer in the deep network model includes: obtaining the quantization value and floating-point value of the feature map output by each layer in the deep network model; and determining the quantization sensitivity of the layer according to the quantization value and the floating-point value of each layer. This embodiment calculates the quantization sensitivity of each layer in combination with the quantization value and floating-point value of the feature map output by each layer, and can obtain an accurate value of the quantization sensitivity, further ensuring the accuracy of the network model.

In some embodiments, determining the quantization sensitivity of each layer according to the quantization value and the floating point value of each layer includes: calculating each The error value between the quantized value of the feature map output by the layer and the floating-point value is taken as the first value; the mean of the floating-point values of the feature map output by each layer is calculated as the second value; the ratio of the first value to the second value of each layer is calculated as the quantization sensitivity of the layer; wherein the error value includes: mean square error, cosine value, root mean square error, or mean absolute error.

This embodiment obtains the quantization sensitivity of each layer by calculating the ratio of the error value (or the first value) and the mean value (or the second value), which can obtain the quantization sensitivity relatively quickly, further improving the quantization efficiency of the network model. In addition, in this implementation, the quantization sensitivity is also obtained based on statistical data, and a more accurate value of the quantization sensitivity can be obtained, and the calculation is simple and easy to implement in this application.

In a second aspect, an embodiment of the present application provides a quantization accuracy selection device for a deep network model, comprising: an acquisition module, used to acquire an input image and a deep network model to be processed, quantize the parameters of the deep network model and set the quantization accuracy of each layer in the quantized deep network model to a first accuracy; a determination module, used to input the input image into the quantized deep network model to obtain a feature map output by each layer of the deep network model, and determine the quantization sensitivity of each layer in the deep network model using the feature map output by each layer of the deep network model; a selection module, used to traverse the layers in the deep network model whose quantization accuracy has not been changed to select a layer with the smallest quantization sensitivity, and change the quantization accuracy of the layer to a second accuracy, which is less than the first accuracy; an inference module, used to perform inference using the deep network model after the quantization accuracy has been changed, and if the inference time is greater than the preset target time, return to the selection module; if the inference time is less than or equal to the preset target time, the deep network model that meets expectations is obtained.

In some embodiments, the determination module includes a first acquisition submodule and a second determination submodule. The first acquisition submodule is used to obtain the quantization value and floating-point value of the feature map output by each layer in the deep network model; the second determination submodule is used to calculate the error value of the quantization value and floating-point value of the feature map output by each layer as the first value; calculate the mean of the floating-point value of the feature map output by each layer as the second value; calculate the ratio of the first value to the second value of each layer as the quantization sensitivity of the layer; wherein the error value includes: mean square error, cosine value, root mean square error, or mean absolute error.

In a third aspect, an embodiment of the present application provides a transplantation system, which includes a data processing device and an electronic device. The data processing device is used to implement a quantization accuracy selection method for a deep network model such as any embodiment of the first aspect when executing a computer program to obtain an expected deep network model. The deep network model is transplanted to the electronic device for use in the electronic device.

In a fourth aspect, an embodiment of the present application provides an electronic device, comprising a processor, wherein the transplantation system transplants a deep network model, such as the transplantation system in the third aspect, to the processor of the electronic device, and the processor processes input data through the onboard deep network model to perform a specific task.

In a fifth aspect, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a data processing device, it implements a method for selecting quantization accuracy of a deep network model as in any embodiment of the first aspect.

In a sixth aspect, an embodiment of the present application provides a computer program product. When the computer program product runs on a data processing device, the data processing device executes a method for selecting quantization accuracy of a deep network model as described in any embodiment of the first aspect.

It should be understood that the beneficial effects of the second to sixth aspects can be found in the relevant description of the embodiment of the first aspect and will not be repeated here.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

FIG1 is a diagram of a transplantation system of a deep network model provided by the present application;

FIG2 is a schematic diagram of an implementation flow of a method for selecting quantization accuracy of a deep network model provided in an embodiment of the present application;

FIG3 is a schematic diagram of a process of step S120 provided in an embodiment of the present application;

FIG. 4 is a process diagram of step S122 provided in an embodiment of the present application;

FIG5 is a schematic diagram of the structure of a device for selecting quantization accuracy of a deep network model provided by an embodiment of the present application;

FIG6 is a schematic diagram of the structure of a determination module provided in an embodiment of the present application.

DETAILED DESCRIPTION

In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.

The term "and/or" as used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

"One embodiment" or "some embodiments" described in the specification of this application means that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment. Therefore, the sentences "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

In addition, in the description of the present application, "plurality" means two or more. The terms "first" and "second" are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

In order to illustrate the technical solution described in this application, a specific embodiment is provided below for illustration.

FIG1 is a diagram of a transplantation system for a deep network model provided in the present application, the system comprising an electronic device and a data processing device, the data processing device is used to process the deep network model according to the quantization accuracy selection method provided in one or more embodiments of the present application to obtain a quantized deep network model; the quantized model is transplanted to an electronic device for the user to perform specific tasks through the electronic device, such as image processing, speech recognition, liveness detection, and face recognition. Specifically, the data processing device includes but is not limited to electronic devices with computing capabilities such as computers, tablet computers, servers, or wearable devices, and the server includes but is not limited to independent servers or cloud servers, etc.; the electronic devices include depth cameras, mobile devices, payment terminals, and robots, etc.

In one embodiment, the deep network model is composed of multiple layers, which can be functionally divided into an input layer, a hidden layer, and an output layer, wherein the input layer is used to receive input data, the hidden layer is used to perform a series of nonlinear changes and feature extraction on the input data in order to better complete a specific task, and the output layer is used to output the result of performing a specific task. It should be noted that this application does not limit the number of input layers, hidden layers, and output layers, which can be multiple layers or one layer, depending on the specific task performed by the deep network model.

In this embodiment, the depth in the deep network model refers to the number of layers of the network model. Compared with the traditional neural network, the number of layers is greater and has powerful feature extraction and representation capabilities, but it also means that the deep network model requires higher inference costs and larger storage. The reasoning cost and storage of low-depth network models need to be quantized during the deployment of deep network models, and quantization will lose the accuracy of the model. Based on this, the present application provides a method for selecting the quantization accuracy of a deep network model, which reduces the reasoning cost and storage consumption of the deep network model while ensuring high quantization accuracy, so that the deep network model can also run on devices with limited resources and capabilities such as low computing power and low memory.

FIG2 is a schematic diagram of an implementation flow of a method for selecting quantization accuracy of a deep network model provided in an embodiment of the present application, the method comprising steps S110 to S140.

S110, obtaining an input image and a deep network model to be processed, quantizing parameters of the deep network model and setting the quantization accuracy of each layer in the quantized deep network model to a first accuracy.

In one embodiment, the input image may include one or more images, which may be face images, body images, gesture images, or other images, and the specific content of the image depends on the specific task performed by the deep network model.

In one embodiment, the parameters in the deep network model are high-precision floating-point numbers. In order to reduce the size of the model, increase the speed of inference, and reduce memory and power consumption, they need to be converted into low-precision fixed-point numbers, that is, the parameters of the deep network model are quantized and the quantization accuracy of each layer of the quantized deep network model is the first accuracy (denoted as Q1). Correspondingly, the quantization bit width of each layer of the quantized deep network model is the first bit width. In some possible implementations, the first bit width can be 32 bits, for example.

S120, inputting the input image into the quantized deep network model to obtain a feature map output by each layer of the deep network model, and using the feature map output by each layer of the deep network model to determine the quantization sensitivity of each layer in the deep network model.

Among them, quantization sensitivity indicates the sensitivity of each layer in the deep network model to the approximation of the original floating-point operations with fixed-point operations; the greater the quantization sensitivity of a layer in the deep network model, the more sensitive the layer is to quantization; conversely, the smaller the quantization sensitivity of a layer in the deep network model, the less sensitive the layer is to quantization.

In some embodiments of the present application, as shown in Figure 3, step S120 may include: S121, obtaining the quantization value and floating-point value of the feature map output by each layer in the deep network model; S122, determining the quantization sensitivity of the layer based on the quantization value and floating-point value of each layer.

Assume that the feature map output by the j-th layer in the deep network model is F _j , obtain the quantization value F _{quan tj} and the floating-point value F _floatj of the feature map F _j output by the j-th layer based on the feature map; calculate the quantization sensitivity of the j-th layer according to the quantization value F _{quan tj} and the floating-point value F _floatj of the feature map F _j output by the j-th layer.

This embodiment combines the quantization value and the floating-point value to determine the quantization sensitivity, and subsequently selects a layer with a smaller quantization sensitivity to use a low bit width for quantization, so as to reduce the inference time of the network model, and repeats the cycle until the inference time meets the target value, thereby improving the efficiency of quantization while ensuring the network accuracy.

As a possible implementation method of an embodiment of the present application, as shown in Figure 4, step S122 may include: S1221, calculating the error value between the quantization value and the floating-point value of the feature map output by each layer as the first numerical value; S1222, calculating the mean of the floating-point values of the feature map output by each layer as the second numerical value; S1223, calculating the ratio of the first numerical value to the second numerical value of each layer as the quantization sensitivity of the layer.

In some examples of the present application, the error value between the quantized value and the floating point value may include but is not limited to the mean square error, cosine value, root mean square error, or mean absolute error, etc., preferably the mean square error is used as the error value. Specifically, the quantized value of the feature map F _j output by the jth layer of the deep network model is F _{quan tj} , the floating point value is F _floatj , and the first value is equal to mse(F _{quan tj} , F _floatj ), where, The second value is equal to That is, the mean value of F _floatj , then the quantization sensitivity of the jth layer is This embodiment uses the statistical mean square error to measure the quantization sensitivity, which can obtain a more accurate value of the quantization sensitivity and further ensure the network model accuracy.

S130, traverse the layers in the deep network model whose quantization precision has not been changed to select a layer with the smallest quantization sensitivity, and change the quantization precision of the layer to a second precision; wherein the second precision is less than the first precision.

In some embodiments of the present application, the second precision can be recorded as Q2, Q2 < Q1, and correspondingly, its quantization bit width is the second bit width. In some possible implementations, the second bit width can be, for example, 16 bits or 8 bits. This embodiment selects layers with less quantization sensitivity to use low bit width for quantization, that is, changes these layers with less quantization sensitivity from the first precision to the second precision, thereby reducing the inference time of the network model.

S140, use the deep network model with changed quantization accuracy to perform inference. If the inference time is greater than the preset target time, return to step S130; if the inference time is less than or equal to the preset target time, the process ends and a deep network model that meets expectations is obtained.

In an embodiment of the present application, the quantization sensitivity of each layer is first determined, and then the layer with the smallest quantization sensitivity is selected from the layers whose quantization accuracy has not been changed, and quantization is performed using a low bit width. This cycle is repeated until the inference time meets the preset target time. The process is terminated to obtain a deep network model that meets the expectations, thereby improving the efficiency of quantization while ensuring the accuracy of the network model.

It should be noted that, in some other embodiments, when the reasoning time is equal to the preset target time, the step of returning to step S130 may also be executed. Those skilled in the art may make a choice based on actual conditions, and the embodiments of the present application do not impose specific restrictions on this.

It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

An embodiment of the present application provides a quantization accuracy selection device for a deep network model. For details not described in the quantization accuracy selection device, please refer to the relevant description in the above-mentioned quantization accuracy selection method embodiment, which will not be repeated here.

FIG5 is a schematic diagram of the structure of a quantization precision selection device for a deep network model provided by an embodiment of the present application. The quantization precision selection device includes: an acquisition module 41, which is used to acquire an input image and a deep network model to be processed, quantize the parameters of the deep network model and set the quantization precision of each layer in the quantized deep network model to be the first precision; a determination module 42, which is used to input the input image into the quantized deep network model to obtain the feature map output by each layer of the deep network model, and determine the quantization sensitivity of each layer in the deep network model using the feature map output by each layer of the deep network model; a selection module 43, which is used to traverse the layers in the deep network model whose quantization precision has not been changed to select a layer with the smallest quantization sensitivity, and change the quantization precision of the layer to the second precision, which is less than the first precision; an inference module 44, which is used to perform inference using the deep network model after the quantization precision is changed, and if the inference time is greater than the preset target time, return to the selection module 43; if the inference time is less than or equal to the preset target time, the deep network model that meets the expectations is obtained.

In some embodiments, as shown in FIG6 , the determination module 42 includes a first acquisition submodule 421 and a second determination submodule 422. The first acquisition submodule 421 is used to obtain the quantization value and floating point value of the feature map output by each layer in the deep network model. The second determination submodule 422 is used to determine the quantization sensitivity of each layer according to the quantization value and floating point value of the layer.

In some embodiments, the second determination submodule is specifically used to: calculate the error value between the quantization value and the floating-point value of the feature map output by each layer as the first numerical value; calculate the mean of the floating-point values of the feature map output by each layer as the second numerical value; calculate the ratio of the first numerical value to the second numerical value of each layer as the quantization sensitivity of the layer.

An embodiment of the present application also provides an electronic device, which may include one or more processors. A deep network model is transplanted to the processor of the electronic device through a transplantation system as shown in FIG1 , so that the processor can process input data through the carried deep network model to perform a specific task.

In one embodiment, the electronic device further includes a camera, which is used to capture a two-dimensional image or a depth image and transmit it to a processor; the processor processes the input image based on a deep network model to perform specific tasks, such as face recognition, liveness detection, or payment tasks. It should be noted that the camera and the electronic device can be designed as an integrated whole or independently, and the two communicate data via wired or wireless means, which is not limited here.

In one embodiment, the processor may be a central processing unit (CPU), or other general-purpose processors, neural network processing chips, digital signal processors (DSP), application specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc.

Those skilled in the art will appreciate that the above are merely examples of electronic devices and do not constitute a limitation on electronic devices. Electronic devices may include more or fewer components than shown in the figure, or combine certain components, or different components. For example, electronic devices may also include input and output devices, network access devices, buses, etc.

The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

An embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed, the steps in the embodiment of the method for selecting quantization accuracy of a deep network model can be implemented.

An embodiment of the present application provides a computer program product. When the computer program product is run, the steps in the embodiment of the method for selecting quantization accuracy of a deep network model can be implemented.

In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

In the embodiments provided in the present application, it should be understood that the disclosed devices/electronic devices and methods can be implemented in other ways. For example, the device/electronic device embodiments described above are merely schematic. For example, the division of modules or units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

If the integrated modules/units are implemented as software functional units and sold or used as independent products, they can be stored in a Computer-readable storage medium. Based on such an understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, and the computer program that can be completed by instructing the relevant hardware through a computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device that can carry computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electric carrier signals and telecommunication signals.

The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims (10)

A method for selecting quantization accuracy of a deep network model, characterized by comprising: S110, obtaining an input image and a deep network model to be processed, quantizing parameters of the deep network model and setting the quantization accuracy of each layer in the quantized deep network model to a first accuracy; S120, inputting the input image into the quantized deep network model to obtain a feature map output by each layer of the deep network model, and determining the quantization sensitivity of each layer in the deep network model using the feature map output by each layer of the deep network model; S130, traversing the layers of the deep network model whose quantization precision has not been changed to select a layer with the smallest quantization sensitivity, and changing the quantization precision of the layer to a second precision, where the second precision is less than the first precision; S140, performing reasoning using the deep network model after changing the quantization accuracy, and if the reasoning time is less than or equal to the preset target time, then the process ends to obtain a deep network model that meets expectations. The quantization accuracy selection method according to claim 1, characterized in that the determining the quantization sensitivity of each layer in the deep network model using the feature map output by each layer of the deep network model comprises: Obtaining quantized values and floating-point values of feature maps output by each layer in the deep network model; The quantization sensitivity of each layer is determined according to the quantization value and the floating point value of the layer. The quantization accuracy selection method according to claim 2, characterized in that the determining the quantization sensitivity of each layer according to the quantization value and the floating-point value of the layer comprises: Calculate the error value between the quantized value and the floating point value of the feature map output by each layer as the first value; Calculate the mean of the floating-point values of the feature map output by each layer as the second value; The ratio of the first value to the second value of each layer is calculated as the quantization sensitivity of the layer. The quantization accuracy selection method as described in claim 3 is characterized in that the error value includes: mean square error, cosine value, root mean square error, or mean absolute error. The quantization accuracy selection method according to claim 4 is characterized in that when the error value is the mean square error, the quantization sensitivity calculation method is: assuming that the quantization value of the feature map F _j output by the jth layer of the deep network model is F _{quan tj} , the floating point value is F _floatj , and the first value is equal to mse(F _{quan tj} , F _floatj ), wherein, The second value is equal to That is, the mean value of F _floatj , then the quantization sensitivity of the jth layer is The quantization accuracy selection method according to any one of claims 1 to 5 is characterized in that after performing reasoning using the neural network model after changing the quantization accuracy, it also includes: if the reasoning time is greater than the preset target time, returning to execute step S130. A device for selecting quantization accuracy of a deep network model, characterized by comprising: An acquisition module, used to acquire an input image and a deep network model to be processed, quantize the parameters of the deep network model and set the quantization accuracy of each layer in the quantized deep network model to be a first accuracy; A determination module, configured to input the input image into the quantized deep network model to obtain a feature map output by each layer of the deep network model, and determine the quantization sensitivity of each layer in the deep network model using the feature map output by each layer of the deep network model; A selection module is used to traverse the layers of the deep network model whose quantization precision is not changed to select a layer with the smallest quantization sensitivity, and changing the quantization accuracy of the layer to a second accuracy, where the second accuracy is less than the first accuracy; The inference module is used to perform inference using the deep network model after the quantization accuracy is changed. If the inference time is less than or equal to the preset target time, the deep network model that meets the expectations is obtained. A deep network model transplantation system, characterized in that it includes a data processing device and an electronic device, wherein the data processing device is used to execute the quantization accuracy selection method of the deep network model as described in any one of claims 1 to 6 to obtain a deep network model that meets the expectations, and the deep network model is transplanted to the electronic device for use by the electronic device. An electronic device, comprising a processor, characterized in that the transplantation system according to claim 8 transplants the deep network model to the processor of the electronic device, and the processor processes input data by carrying the deep network model to perform a specific task. A computer storage medium storing a computer program, wherein the computer program, when executed, implements the quantization accuracy selection method as described in any one of claims 1 to 6.