Model processing device networking method and model processing system

Abstract

The application relates to the field of machine learning or deep learning, and provides a networking method of a model processing device and a model processing system. The method comprises the following steps: when there is only one model processing device, configuring the only one model processing device as a first center machine Mc-1; when there are multiple model processing devices, configuring one of the multiple model processing devices as a second center machine Mc-2, and configuring the remaining model processing devices except the one configured as the second center machine Mc-2 as edge devices; when the only one model processing device is configured as the first center machine Mc-1, the first center machine Mc-1 is used for model training and model inference; when one of the multiple model processing devices is configured as the second center machine Mc-2, the second center machine Mc-2 is used for model training, and the edge devices are used for model inference. The technical scheme of the application can reduce the cost of a production line and reduce the adverse effect on the production line.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of machine learning or deep learning, in particular to a networking method of a model processing device and a model processing system. BACKGROUND

[0002] With the development of machine automation and the long-term large amount of manual input in a large number of industrial production lines, deep learning applied to industrial production and finished product detection has become a general trend. By training a model to generate a quality detection model for a sample, and then using the trained model to detect and classify the quality defects of the products in the production process according to the detection requirements, the detection speed and accuracy are improved while saving manual detection.

[0003] However, in different factory environments, due to the differences in the number and requirements of production lines, the management and maintenance of production lines and the auditing of detection data have always been a pain point in the industry. In some scenarios, the redundancy of model processing devices relative to production lines will lead to an increase in management and maintenance costs, and in other scenarios, it will bring repetitive operation costs of models and data. Repetitive operations on multiple edge devices inevitably lead to misoperations or omissions, and also affect the immediate effectiveness and mass production capacity of the production line. SUMMARY

[0004] The present application provides a networking method of a model processing device and a model processing system, which can reduce the cost of the production line and reduce the adverse effects on the production line.

[0005] In one aspect, the present application provides a networking method of a model processing device, the method comprising:

[0006] When there is only one model processing device, the only one model processing device is configured as a first center machine Mc-1;

[0007] When there are multiple model processing devices, one of the multiple model processing devices is configured as a second center machine Mc-2, and the remaining model processing devices except the second center machine Mc-2 are configured as edge devices;

[0008] When the only one model processing device is configured as the first center machine Mc-1, the first center machine Mc-1 is used for model training and model inference, and when one of the multiple model processing devices is configured as the second center machine Mc-2, the second center machine Mc-2 is used for model training, and the edge device is used for model inference.

[0009] In another aspect, the present application provides a model processing system, the system comprising:

[0010] only one model processing device is configured as the first central machine Mc-1; or

[0011] a plurality of model processing devices, one of the plurality of model processing devices is configured as a second central machine Mc-2, and the rest of the model processing devices except the one configured as the second central machine Mc-2 are configured as edge devices;

[0012] When only one model processing device is configured as the first central machine Mc-1, the first central machine Mc-1 is used for model training and model inference, and when one of the plurality of model processing devices is configured as the second central machine Mc-2, the second central machine Mc-2 is used for model training, and the edge devices are used for model inference.

[0013] From the above technical solutions provided by the present application, when there is only one model processing device, the only one model processing device is configured as the first central machine Mc-1; when there are a plurality of model processing devices, one of the plurality of model processing devices is configured as the second central machine Mc-2, and the rest of the model processing devices except the one configured as the second central machine Mc-2 are configured as edge devices. On the one hand, when only one model processing device is configured as the first central machine Mc-1, the first central machine Mc-1 is used for model training and model inference. For a small production line, only one model processing device is needed to complete all detection processes related to industrial automation, and the configuration is not only not redundant, but also can save a lot of management and maintenance costs. On the other hand, when one of the plurality of model processing devices is configured as the second central machine Mc-2, and the rest of the model processing devices are configured as edge devices, the second central machine Mc-2 is used for model training, and the edge devices are used for model inference, which can reduce the repetitive operation of models and data, not only objectively reduce the probability of errors caused by repetitive operation, thereby reducing costs, but also reduce various adverse effects on the production line. BRIEF DESCRIPTION OF DRAWINGS

[0014] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can also obtain other drawings according to these drawings without creative labor.

[0015] Figure 1 is a flowchart of the networking method of the model processing device provided by the embodiments of the present application;

[0016] Figure 2 is a configuration information schematic diagram of each node of the tree-shaped node after the networking of the seven edge devices provided by the embodiments of the present application;

[0017] Figure 3 is a schematic diagram of a networking scheme of a model processing device provided by an embodiment of the present application.

[0018] Figure 4 is a structural schematic of the device. DETAILED DESCRIPTION

[0019] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0020] In this specification, adjectives such as first and second can be used merely to differentiate one element or action from another element or action, without necessarily implying any actual such relationship or order. Where the context permits, reference to an element or component or step (etc.) by the indefinite article "a" or "an" does not exclude the existence of, or the possibility of using, more than one of the element, component or step (etc.).

[0021] In this specification, for the convenience of description, the sizes of the various parts shown in the drawings are not drawn in accordance with the actual proportional relationship.

[0022] As shown in the accompanying Figure 1 is a flowchart of a networking method of a model processing device proposed by the present application. Figure 1 The example method can be applied to any real-time detection system that needs to independently or batch manage the quality of production line finished products. The system after the networking of the model processing device is a production line management system that can integrate a central machine for independent training and reasoning or a model training, model inference, and production line data management function of the production line management system containing one or more edge devices. Figure 1 The example method mainly includes steps S101 and S102, which are described in detail as follows:

[0023] Step S101: When there is only one model processing device, the only model processing device is configured as a first central machine Mc-1.

[0024] In the embodiments of the present application, the model processing device is various electronic devices capable of training and / or making inference of a model, including ordinary computers, servers, etc. Generally, model training and model inference are performed on two physically separated devices. In other words, at least two devices are required to perform model training and model inference. In the embodiments of the present application, when there is only one model processing device, the only model processing device is configured as a center machine (to distinguish from devices configured as center machines in other scenarios, it is referred to as a first center machine Mc-1 here). When the only model processing device is configured as the first center machine Mc-1, the first center machine Mc-1 is used for model training and model inference. For a small production line, configuring a model processing device as a center machine that can be used for both model training and model inference means that only one device is required to complete all detection processes related to industrial automation, thereby reducing device redundancy and reducing production and operation costs of the production line.

[0025] Step S102: When there are multiple model processing devices, one of the multiple model processing devices is configured as a second center machine Mc-2, and the remaining model processing devices other than the second center machine Mc-2 are configured as edge devices.

[0026] For large and medium-sized production lines, a single model processing device generally cannot meet the requirements and multiple model processing devices may be required. When there are multiple model processing devices, one of the multiple model processing devices is configured as a center machine (to distinguish from devices configured as center machines in other scenarios, it is referred to as a second center machine Mc-2), and the remaining model processing devices other than the second center machine Mc-2 are configured as edge devices. As mentioned above, model training and model inference are generally performed on two physically separated devices. Therefore, when one of the multiple model processing devices is configured as the second center machine Mc-2, the second center machine Mc-2 can be used for model training, and the edge devices can be used for model inference, for example, the second center machine Mc-2 obtains a data set used for training, such as an image, from the edge devices, and after the model is trained, the trained model is distributed to the edge devices, and the edge devices use the trained model for inference.

[0027] In an embodiment of the present application, when there are multiple model processing devices, any one of the edge devices can be configured as the center machine according to the needs of the production line, and the second center machine Mc-2 is configured as an edge device, which means that for the networking method of the model processing device of the present application, the center machine and the edge device can be exchanged according to the needs of the actual production line or in order to realize multi-layer management and statistics of the production line model and data. The direct benefit is that the data can be backed up bidirectionally, and the operation is convenient when managing multiple devices. For example, assuming that there are 8 model processing devices in the first level of the networking system, any one of the model processing devices can be defined as a center machine, then on any one of the model processing devices, the production data of the other seven model processing devices can be accessed, and the production management in this system can also be managed through any one of the model processing devices to manage the production and data review of the other seven model processing devices. In other words, if the 8 model processing devices deployed in multiple floors (even buildings) need to be maintained by the same person, the operation can be completed on the nearest model processing device without the need for hardware reboot.

[0028] When one of the multiple model processing devices is configured as the second center machine Mc-2, and the remaining model processing devices are configured as edge devices, these edge devices can be single devices (for example, two model processing devices except for the one configured as the second center machine Mc-2, only one model processing device is configured as an edge device, which is a single device), belong to a group or not belong to any group. Figure 2 As shown in the figure, it is a configuration information diagram of each node of the tree-shaped nodes after the networking of 7 edge devices, Edge machine 1, Edge machine 2, Edge machine 3, Edge machine A, Edge machine B, Edge machine i and Edge machine ii, etc. 7 edge devices belong to two groups, that is, Production Group 1 and Production Group 2 in the figure, among which, Edge machine 1, Edge machine 2 and Edge machine 3 belong to the group Production Group 1, Edge machine A and Edge machine B belong to the group Production Group 2, and Edge machine i and Edge machine ii do not belong to any group. Figure 2In the embodiment, the first-level node under the node remotes_group is the group name of the group to which the edge device belongs, the first-level node under the group to which the edge device belongs (i.e., the second-level node under the node remotes_group) is the name of the edge device, and the first-level node under the node remotes is the name of the edge device, wherein the device-related information of the edge device includes but is not limited to IP, login username, connection state, connection protocol, automatic connection setting, and last connection time, etc. It should be noted that, Figure 2 The edge devices under the node remotes, i.e., the edge devices Edgemachine i and Edge machine ii that have not been grouped, can be incorporated into a certain group at any time, or can remain ungrouped for use, and the edge devices that belong to a certain group can also be removed from the group to become devices that do not belong to any group.

[0029] In the embodiment, the center machine and all edge devices can be respectively configured with connection information, and the center machine can perform independent or batch model distribution and rollback management on all connected edge devices. When the edge device belongs to a group, it means that different grouping strategies can be configured according to the needs of the real production line, and model distribution and inference control can be performed on a single edge device in the group or on all edge devices in the same group. Therefore, the technical solution of the present application supports grouping batch operation, and operations related to the production line can be managed and operated individually or in batches, including model deployment, model inference, production line production management, and production line data management, etc., which can reduce the cost of the production line and improve the efficiency of the production line.

[0030] In one embodiment of the present application, when the edge device belongs to a group, the upper-level node of the group to which the edge device belongs is configured as a third center machine Mc-2, and further, if the third center machine Mc-2 has an upper-level node, the third center machine Mc-2 can be configured as an edge device, which means that a certain model processing device can be configured as a center machine that has at least one edge device, and the center machine can also be configured as a certain edge device. The following describes an example networking scheme. Figure 3

[0031] In Figure 3 ​In the networking scheme of the example model processing device, the center machine (in the embodiments of the present application, the center machine can be referred to as an “integrated management device”; in order to distinguish from other center machines, it is referred to as “center machine 1” here) 1 is the total center machine of the whole system, and manages model deployment, inference detection and data statistics of all devices. The edge device 1 and the edge device 2 are edge devices of the production line, and the model deployment and the data statistics can only be managed and controlled by the center machine 1. The edge device 3 can be configured as an edge device of the production line, and the model deployment and the data statistics can only be managed and controlled by the center machine 1, and at the same time, it can also be configured as a center machine (in order to distinguish from other center machines, it is referred to as “center machine 2” here) for managing and controlling model deployment and data statistics of the edge devices 5 to 8. The edge device 4 can be configured as an edge device of the production line, and the model deployment and the data statistics can only be managed and controlled by the center machine 1, and at the same time, it can also be configured as a center machine (in order to distinguish from other center machines, it is referred to as “center machine 3” here) for managing and controlling model deployment and data statistics of the edge devices 9 to 20. The model deployment, inference detection and data statistics of the edge devices 5 to 8 can be managed and controlled by the center machine 1 and the center machine 2. The edge device 9 can be configured as an edge device of the production line, and the model deployment and the data statistics can be managed and controlled by the center machine 1 and the center machine 3, and at the same time, it can also be configured as a center machine (in order to distinguish from other center machines, it is referred to as “center machine 4” here) for managing and controlling model deployment and data statistics of the edge devices 12 to 15. The edge device 10 can be configured as an edge device of the production line, and the model deployment and the data statistics can be managed and controlled by the center machine 1 and the center machine 4, and at the same time, it can also be configured as a center machine (in order to distinguish from other center machines, it is referred to as “center machine 5” here) for managing and controlling model deployment and data statistics of the edge devices 16 and 17. The edge device 11 can be configured as an edge device of the production line, and the model deployment and the data statistics can be managed and controlled by the center machine 1 and the center machine 4, and at the same time, it can also be configured as a center machine (in order to distinguish from other center machines, it is referred to as “center machine 6” here) for managing and controlling model deployment and data statistics of the edge devices 18 to 20. The model deployment, inference detection and data statistics of the edge devices 12 to 15 can be managed and controlled by the center machine 1, the center machine 3 and the center machine 4. The model deployment, inference detection and data statistics of the edge devices 16 and 17 can be managed and controlled by the center machine 1, the center machine 3 and the center machine 5. The model deployment, inference detection and data statistics of the edge devices 18 to 20 can be managed and controlled by the center machine 1, the center machine 3 and the center machine 6.

[0032] By Figure 3The networking scheme of the example model processing device can support multi-layer network management, including various model training and model inference, and data collection and review of various product lines. For example, the first layer of the network reviews the entire factory output, the second layer of the network reviews the building output, the third layer reviews the A product output in the building, the fourth layer reviews the A product b model output in the building, and so on.

[0033] In the above embodiment, when there are multiple model processing devices and one of the multiple model processing devices is configured as the second center machine Mc-2, the remaining model processing devices except the second center machine Mc-2 are configured as edge devices. The second center machine Mc-2 or the third center machine Mc-3 obtains data generated in the production process from the edge devices, classifies and counts the data into a database and a local file, and presents the data to the user for operation according to the group to which the edge devices belong or the individual edge devices, including data statistics, review, rollback test, and the like.

[0034] From the above description of the accompanying drawings, Figure 1 The networking method of the example model processing device can be known that when there is only one model processing device, the only one model processing device is configured as the first center machine Mc-1; when there are multiple model processing devices, one of the multiple model processing devices is configured as the second center machine Mc-2, and the remaining model processing devices except the second center machine Mc-2 are configured as edge devices. On the one hand, when the only one model processing device is configured as the first center machine Mc-1, the first center machine Mc-1 is used for model training and model inference. For a small production line, only one model processing device is needed to complete all detection processes related to industrial automation, and the configuration is not only not redundant, but also can save a lot of management and maintenance costs. On the other hand, when one of the multiple model processing devices is configured as the second center machine Mc-2 and the remaining model processing devices are configured as edge devices, the second center machine Mc-2 is used for model training, and the edge devices are used for model inference, which can reduce the repetitive operation of models and data, not only objectively reduce the probability of errors caused by repetitive operation, thereby reducing costs, but also reduce various adverse effects on the production line.

[0035] The embodiments of the present application also provide a model processing system, which comprises: only one model processing device configured as a first central machine Mc-1; or a plurality of model processing devices, wherein one of the plurality of model processing devices is configured as a second central machine Mc-2, and the remaining model processing devices of the plurality of model processing devices except the one configured as the second central machine Mc-2 are configured as edge devices; when the only one model processing device is configured as the first central machine Mc-1, the first central machine Mc-1 is used for model training and model inference, and when one of the plurality of model processing devices is configured as the second central machine Mc-2, the second central machine Mc-2 is used for model training, and the edge devices are used for model inference.

[0036] From the above model processing system, on the one hand, when the only one model processing device is configured as the first central machine Mc-1, the first central machine Mc-1 is used for model training and model inference, and for a small production line, only one model processing device is needed to complete all detection processes related to industrial automation, and the configuration is not only not redundant, but also can save a lot of management and maintenance costs; on the other hand, when one of the plurality of model processing devices is configured as the second central machine Mc-2, and the remaining model processing devices are configured as edge devices, the second central machine Mc-2 is used for model training, and the edge devices are used for model inference, which can reduce repetitive operations of models and data, not only objectively reduce the probability of errors caused by repetitive operations, thereby reducing costs, but also reduce various adverse effects on the production line.

[0037] Optionally, in the model processing system of the above example, the edge devices are single devices, belong to a group, or do not belong to any group.

[0038] Optionally, in the model processing system of the above example, when the edge devices belong to a group, the model processing device corresponding to the upper node of the group to which the edge devices belong is configured as a third central machine Mc-2.

[0039] Optionally, in the model processing system of the above example, if the third central machine Mc-2 has an upper node, the third central machine Mc-2 is configured as an edge device.

[0040] Optionally, in the model processing system of the above example, when there are a plurality of model processing devices, according to the needs of the production line, any one of the edge devices is configured as a central machine, and the second central machine Mc-2 is configured as an edge device.

[0041] Optionally, in the model processing system of the above example, the second central machine Mc-2 or the third central machine Mc-3 can be used to obtain data generated in the production process from the edge devices, classify and count the data into a database and a local file, and then present the data to a user for operation according to the group to which the edge devices belong or according to the single edge device.

[0042] This application embodiment also provides a networking device for a model processing device, which may include a first configuration module and a second configuration module, wherein:

[0043] The first configuration module is used to configure the only model processing device as the first central machine Mc-1 when there is only one model processing device.

[0044] The second configuration module is used to configure one of the multiple model processing devices as the second central machine Mc-2, and the other model processing devices besides the one configured as the second central machine Mc-2 as edge devices when there are multiple model processing devices; when only one model processing device is configured as the first central machine Mc-1, the first central machine Mc-1 is used for model training and model inference; when one of the multiple model processing devices is configured as the second central machine Mc-2, the second central machine Mc-2 is used for model training and the edge devices are used for model inference.

[0045] Figure 4 This is a schematic diagram of the structure of a device provided in one embodiment of this application. For example... Figure 4 As shown, the device 4 in this embodiment mainly includes: a processor 40, a memory 41, and a computer program 42 stored in the memory 41 and executable on the processor 40, such as a program for a networking method of a model processing device. When the processor 40 executes the computer program 42, it implements the steps in the above-described network method embodiment of the model processing device, for example... Figure 1 The steps S101 and S102 are shown. Alternatively, when the processor 40 executes the computer program 42, it implements the functions of each module / unit in the above-described device embodiments, such as the functions of the first configuration module and the second configuration module.

[0046] Exemplarily, the computer program 42 of the networking method of the model processing device mainly includes: when there is only one model processing device, configuring the only one model processing device as a first center machine Mc-1; when there are multiple model processing devices, configuring one of the multiple model processing devices as a second center machine Mc-2, and configuring the remaining model processing devices of the multiple model processing devices except the one configured as the second center machine Mc-2 as edge devices; when the only one model processing device is configured as the first center machine Mc-1, the first center machine Mc-1 is used for model training and model inference, and when one of the multiple model processing devices is configured as the second center machine Mc-2, the second center machine Mc-2 is used for model training, and the edge devices are used for model inference. The computer program 42 can be divided into one or more modules / units, one or more modules / units are stored in the memory 41 and executed by the processor 40 to complete the present application. One or more modules / units can be a series of computer program instruction segments capable of completing a specific function, which is used to describe the execution process of the computer program 42 in the device 4. For example, the computer program 42 can be divided into the functions of the first configuration module and the second configuration module (modules in the virtual device), and the specific functions of each module are as follows: the first configuration module is used for configuring the only one model processing device as the first center machine Mc-1 when there is only one model processing device; the second configuration module is used for configuring one of the multiple model processing devices as the second center machine Mc-2 when there are multiple model processing devices, and configuring the remaining model processing devices of the multiple model processing devices except the one configured as the second center machine Mc-2 as edge devices; when the only one model processing device is configured as the first center machine Mc-1, the first center machine Mc-1 is used for model training and model inference, and when one of the multiple model processing devices is configured as the second center machine Mc-2, the second center machine Mc-2 is used for model training, and the edge devices are used for model inference.

[0047] The device 4 can include but is not limited to the processor 40 and the memory 41. Those skilled in the art can understand that, Figure 4 The device 4 is only an example and does not constitute a limitation on the device 4, and can include more or fewer components than the illustration, or combine certain components, or different components, for example, the computing device can also include an input / output device, a network access device, a bus, etc.

[0048] The processor 40 can be a central processing unit (CPU), and can also be other general-purpose processors, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic device, discrete hardware component, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor.

[0049] The memory 41 can be an internal storage unit of the device 4, such as a hard disk or a memory of the device 4. The memory 41 can also be an external storage device of the device 4, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. Further, the memory 41 can include both the internal storage unit and the external storage device of the device 4. The memory 41 is used to store computer programs and other programs and data required by the device. The memory 41 can also be used to temporarily store data that has been output or will be output.

[0050] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above functional units and modules is exemplified, and in actual application, the above functions can be completed by different functional units and modules according to needs, that is, the internal structure of the apparatus is divided into different functional units or modules to complete all or part of the above described functions. Each functional unit and module in the embodiment can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit, and the integrated unit can be realized in the form of hardware or in the form of software functional unit. In addition, the specific names of each functional unit and module are only for easy distinction, and do not limit the protection scope of the present application. The specific working process of the units and modules in the apparatus can refer to the corresponding process in the foregoing method embodiments, which will not be described here.

[0051] In the above embodiments, the description of each embodiment has its own emphasis, and the parts not described or recorded in detail in a certain embodiment can be referred to the related description of other embodiments.

[0052] Those skilled in the art can understand that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0053] In the embodiments provided in the present application, it should be understood that the disclosed apparatuses / devices and methods can be implemented in other ways. For example, the above-described apparatus / device embodiments are merely illustrative, for example, the division of modules or units is merely a logical function division, and there can be another division manner in actual implementation, for example, a plurality of units or components can be combined or integrated into another device, or some features can be ignored or not executed. In addition, the displayed or discussed mutual coupling or direct coupling or communication connection can be indirect coupling or communication connection through some interface, device or unit, and can be electrical, mechanical or other forms.

[0054] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, that is, they can be located in one place, or can be distributed on a plurality of network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiments.

[0055] In addition, each functional unit in each embodiment of the present application can be integrated into a processing unit, or each unit can exist physically independently, or two or more units can be integrated into one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0056] The integrated modules / units, if implemented in the form of software function units and sold or used as independent products, can be stored in a storage medium. Based on such understanding, all or part of the processes in the above-mentioned embodiment methods can also be completed by a computer program instructing related hardware. The computer program of the networking method of the model processing device can be stored in a storage medium. When executed by a processor, the computer program can implement the steps of each method embodiment described above, that is, when there is only one model processing device, the only one model processing device is configured as the first center machine Mc-1; when there are multiple model processing devices, one of the multiple model processing devices is configured as the second center machine Mc-2, and the remaining model processing devices except the one configured as the second center machine Mc-2 are configured as edge devices; when the only one model processing device is configured as the first center machine Mc-1, the first center machine Mc-1 is used for model training and model inference, and when one of the multiple model processing devices is configured as the second center machine Mc-2, the second center machine Mc-2 is used for model training, and the edge device is used for model inference. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or some intermediate forms, etc. The storage medium can include any entity or device capable of carrying computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electrical carrier signal, telecommunication signal, and software distribution medium, etc. It should be noted that the content contained in the storage 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 storage medium does not include electrical carrier signals and telecommunication signals.

[0057] The above embodiments are only used to illustrate the technical solutions of the present application, but not limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacements for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application. The above specific embodiments further illustrate the purpose, technical solutions and beneficial effects of the present application, and it should be understood that the above description is only the specific embodiment of the present application, and does not limit the protection scope of the present application, and any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A networking method for model processing devices, characterized in that, The method includes: When there is only one model processing device, the only model processing device is configured as the first central machine Mc-1; When there are multiple model processing devices, one of the multiple model processing devices is configured as the second central machine Mc-2, and the other model processing devices are configured as edge devices. When the only model processing device is configured as the first central machine Mc-1, the first central machine Mc-1 is used for model training and model inference. When one of the multiple model processing devices is configured as the second central machine Mc-2, the second central machine Mc-2 is used for model training, and the edge device is used for model inference. When there are multiple model processing devices, any one of the edge devices can be configured as a central machine according to the needs of the production line, and the second central machine Mc-2 can be configured as an edge device. When the edge device belongs to a group, different grouping strategies are configured according to the needs of the actual production line. Model distribution and inference control are performed on individual edge devices within the group, or on all edge devices within the same group.

2. The networking method for the model processing device as described in claim 1, characterized in that, The edge device may be a single device, belong to a group, or not belong to any group.

3. The networking method for the model processing device as described in claim 2, characterized in that, The method further includes: When the edge device belongs to a group, the model processing device corresponding to the upper-level node of the group to which the edge device belongs is configured as the third central machine Mc-2.

4. The networking method for the model processing device as described in claim 3, characterized in that, The method further includes: If the third central machine Mc-2 has a superior node, the third central machine Mc-2 will be configured as an edge device.

5. The networking method for the model processing device as described in any one of claims 1 to 4, characterized in that, The method further includes: The second central machine Mc-2 or the third central machine Mc-3 acquires data generated during the production process from the edge devices; After the data is categorized and statistically analyzed in the database and local files, it is presented to the user for operation according to the group to which the edge device belongs or according to the individual edge device.

6. A model processing system, characterized in that, The system includes: The only model processing device configured as the first central machine, Mc-1; or Multiple model processing devices, one of which is configured as the second central machine Mc-2, and the remaining model processing devices other than the one configured as the second central machine Mc-2 are configured as edge devices; When the only model processing device is configured as the first central machine Mc-1, the first central machine Mc-1 is used for model training and model inference. When one of the multiple model processing devices is configured as the second central machine Mc-2, the second central machine Mc-2 is used for model training, and the edge devices are used for model inference. When there are multiple model processing devices, any one of the edge devices can be configured as a central machine and the second central machine Mc-2 can be configured as an edge device according to the needs of the production line. When the edge devices belong to a group, different grouping strategies are configured according to the needs of the actual production line to perform model distribution and inference control on a single edge device within the group, or to perform model distribution and inference control on all edge devices within the same group.

7. The model processing system as described in claim 6, characterized in that, The edge device may be a single device, belong to a group, or not belong to any group.

8. The model processing system as described in claim 7, characterized in that, When the edge device belongs to a group, the model processing device corresponding to the upper-level node of the group to which the edge device belongs is configured as the third central machine Mc-2.

9. The model processing system as described in claim 8, characterized in that, If the third central machine Mc-2 has a superior node, the third central machine Mc-2 is configured as an edge device.

Images