Algorithm process method, apparatus and device, and storage medium

Abstract

The invention discloses an algorithm process method, apparatus and device, and a storage medium. The method comprises the steps of obtaining power operation inspection function information; processing the electric power operation and maintenance function information based on a large language model to obtain first function module information capable of cooperatively realizing a function corresponding to the electric power operation and maintenance function information in the large language model, the first function module information comprises a first function module name of at least one first function module and a data input format and a data output format of each first function module, and the at least one first function module is contained in the large language model; and determining a first algorithm flow for realizing a function corresponding to the power operation and maintenance function information according to the first function module information. The algorithm process of the power operation and maintenance function is automatically generated by using the large language model, and developers do not need to code codes to generate the algorithm process, so that the labor cost and the time cost are saved, and the efficiency of generating the algorithm process is improved.

Description

Technical Field

[0001] The present application belongs to the field of information processing technology, and in particular, relates to an algorithm flow method, device, equipment and storage medium. Background Art

[0002] There are many algorithms in the field of power operation and inspection, for example, the field includes equipment detection algorithms, defect detection algorithms, image classification algorithms, etc. In the face of different algorithm requirements, developers need to manually write code to build algorithm flows corresponding to different algorithms to meet the algorithm requirements through the algorithm flow. In the face of numerous algorithm requirements for power operation and inspection, the method of generating algorithm flows corresponding to algorithm requirements by developers writing code has the problems of high labor and time costs and low efficiency. Summary of the invention

[0003] The embodiment of the present application provides an implementation scheme different from the prior art to solve the technical problems of high time and labor costs and low efficiency in generating algorithm processes in the prior art.

[0004] In a first aspect, the present application provides an algorithm flow method, applicable to a server, comprising: obtaining power operation and inspection function information; processing the power operation and inspection function information based on a large language model to obtain first function module information in the large language model that can collaboratively implement the function corresponding to the power operation and inspection function information, the first function module information including a first function module name of at least one first function module, a data input format and a data output format of each first function module, and the at least one first function module is included in the large language model; determining a first algorithm flow for implementing the function corresponding to the power operation and inspection function information according to the first function module information.

[0005] In the second aspect, another algorithm flow method provided by the present application is applicable to the client, including: receiving a first algorithm flow sent by a server; relevant technical personnel verifying whether the first algorithm flow can realize the function corresponding to the power operation and inspection function information; if so, sending the first algorithm flow as the second algorithm flow to the server; if not, obtaining information to be processed based on the first algorithm flow and the power operation and inspection function information, wherein the information to be processed includes adding or deleting at least one third function module, and adding or deleting a third directed connection relationship corresponding to the at least one third function module; based on the information to be processed and the first algorithm flow, obtaining the second algorithm flow, and sending the second algorithm flow to the server.

[0006] In a third aspect, an algorithm flow device provided by the present application is suitable for a server, comprising: an acquisition unit, used to acquire power operation and inspection function information; a processing unit, used to process the power operation and inspection function information based on a large language model, and obtain first function module information in the large language model that can collaboratively implement the function corresponding to the power operation and inspection function information, the first function module information including a first function module name of at least one first function module, a data input format and a data output format of each first function module, and the at least one first function module is included in the large language model; a determination unit, used to determine a first algorithm flow for implementing the function corresponding to the power operation and inspection function information according to the first function module information.

[0007] In a fourth aspect, the present application provides an electronic device comprising: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to execute any method in the first aspect, the second aspect, or any possible implementation of the first aspect or the second aspect by executing the executable instructions.

[0008] In a fifth aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the computer program implements any method in the first aspect, the second aspect, or any possible implementation of the first aspect or the second aspect.

[0009] In a sixth aspect, an embodiment of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the method described in the first aspect, the second aspect, or any possible implementation manner of the first aspect or the second aspect.

[0010] The solution provided by the present application is to obtain the power operation and inspection function information; process the power operation and inspection function information based on the large language model to obtain the first function module information in the large language model that can collaboratively realize the function corresponding to the power operation and inspection function information, the first function module information includes the first function module name of at least one first function module, the data input format and the data output format of each first function module, and the at least one first function module is included in the large language model; determine the first algorithm flow scheme for realizing the function corresponding to the power operation and inspection function information according to the first function module information. The algorithm flow of the power operation and inspection function is automatically generated by the large language model, without the need for developers to code to generate the algorithm flow, saving manpower costs and time costs, and improving the efficiency of generating the algorithm flow. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following is a brief introduction to the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0012] Figure 1 A schematic diagram of the structure of a system provided in one embodiment of the present application;

[0013] Figure 2 A flowchart of an algorithm flow method provided in one embodiment of the present application;

[0014] Figure 3 A schematic diagram of a structure of a generation algorithm process provided in one embodiment of the present application;

[0015] Figure 4 A structural diagram of a modification algorithm process provided in an embodiment of the present application;

[0016] Figure 5 A flowchart of another algorithm flow method provided in one embodiment of the present application;

[0017] Figure 6 A schematic diagram of the structure of an algorithm flow device provided in one embodiment of the present application;

[0018] Figure 7 A schematic diagram of the structure of another algorithm flow device provided in an embodiment of the present application;

[0019] Figure 8 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0020] The embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, but cannot be understood as limiting the present application.

[0021] The terms "first" and "second" etc. in the specification, claims and drawings of the embodiments of the present application are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein, for example. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, the process, method, system, product or equipment comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or equipment.

[0022] First, some terms in the embodiments of the present application are explained below to facilitate understanding by those skilled in the art.

[0023] Power operation and inspection: refers to the operation, maintenance and inspection of the power system to ensure the safe, stable and efficient operation of the power system.

[0024] Power operation and maintenance diagram: referred to as operation and maintenance diagram, refers to the chart or drawing of the operation and maintenance of the power system.

[0025] Power line hardware: refers to various metal accessories for installing and fixing power lines, including insulators, tension clamps, suspension clamps, parallel groove clamps, connecting pipes, hanging plates, bolts, cross arms, shock-absorbing hammers, etc.

[0026] There are many algorithms in the field of power operation and inspection, for example, the field of power operation and inspection includes equipment detection algorithms, defect detection algorithms, image classification algorithms, etc. In the face of different algorithm requirements, developers need to manually write code to build algorithm flows corresponding to different algorithms so as to meet the algorithm requirements through the algorithm flow. In the face of numerous algorithm requirements for power operation and inspection, the method of generating algorithm flows corresponding to algorithm requirements by developers writing code has the problems of high labor and time costs and low efficiency. In addition, when developers manually write code to build algorithm flows corresponding to different algorithms, code writing errors, code redundancy and other problems are prone to occur, resulting in a waste of server resources and the problem of low efficiency of the generated algorithm flow.

[0027] The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems are described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0028] Figure 1A schematic diagram of a system structure provided for an exemplary embodiment of the present application, the system includes a server and a client:

[0029] The server is used to obtain power operation and inspection function information; based on the large language model, the power operation and inspection function information is processed to obtain the first function module information in the large language model that can collaboratively implement the function corresponding to the power operation and inspection function information, the first function module information includes the first function module name of at least one first function module, the data input format and the data output format of each first function module, and the at least one first function module is included in the large language model; according to the first function module information, a first algorithm flow for implementing the function corresponding to the power operation and inspection function information is determined.

[0030] The client is used to receive a first algorithm flow sent by a server; relevant technical personnel verify whether the first algorithm flow can realize the function corresponding to the power operation and inspection function information; if so, the first algorithm flow is sent to the server as the second algorithm flow; if not, based on the first algorithm flow and the power operation and inspection function information, the information to be processed is obtained, wherein the information to be processed includes adding or deleting at least one third function module, and adding or deleting a third directed connection relationship corresponding to the at least one third function module; based on the information to be processed and the first algorithm flow, the second algorithm flow is obtained, and the second algorithm flow is sent to the server.

[0031] The execution principles and interaction processes of the various components in the system embodiment, such as the server and the client, can be found in the description of the following method embodiments.

[0032] Figure 2 A flowchart of an algorithm flow method provided for an exemplary embodiment of the present application, the execution subject of the method may be a server, and the method at least includes the following steps S101-S103:

[0033] S101, obtaining power operation and inspection function information;

[0034] Optionally, the power operation and inspection function information refers to information used to describe the power operation and inspection function, and can also be understood as requirement information for an algorithm flow to implement the power operation and inspection function.

[0035] Optionally, the power operation and inspection function refers to the function of implementing maintenance and inspection of the power system.

[0036] S102, processing the electric power operation and inspection function information based on the large language model to obtain first function module information in the large language model that can collaboratively implement the function corresponding to the electric power operation and inspection function information, wherein the first function module information includes a first function module name of at least one first function module, a data input format and a data output format of each first function module, and the at least one first function module is included in the large language model;

[0037] Optionally, the large language model can be a large language model such as ChatGPT, Wenxin Yiyan, Tongyi Qianwen, and Spark Large Model, or a private large language model fine-tuned internally. For details, please refer to the prior art, which will not be described in detail here.

[0038] Optionally, the first function module information is a character string representing process information output by the large language model.

[0039] Optionally, a first functional module refers to a sub-model for realizing a sub-function, and the first functional module information of the function corresponding to the power operation and inspection function information is realized by cooperating with at least one sub-function corresponding to at least one first functional module. Among them, the sub-models include but are not limited to image reading models, detection models, CLIP models, output models, instance segmentation models, OCR models, SAM models, etc., which can be specifically set according to actual needs. For the relevant explanations of the OCR model, SAM model and CLIP model, please refer to the prior art, and no further details will be given here.

[0040] Optionally, the first functional module is a functional module stored in the preset functional module library in the large language model.

[0041] Specifically, the function module is a plurality of function modules stored in the preset function module library. The input data, output data, input data type, output data type and description content of the corresponding achievable functions of the function module are converted into a natural language of a preset format for encapsulation and stored in the preset function module library.

[0042] Exemplarily, the first functional module may also be a model for realizing target segmentation, a model for mark reading, a model for personnel work identification, a model for data indexing, a model for visualization, a model for database indexing, and a model for post-processing logic, etc. It is encapsulated in a preset format through natural language and stored in a preset functional module library.

[0043] Exemplarily: the preset format is: Description: functional description of the function that can be implemented by the functional module, input data: [{0: data, type: data type identifier}, {1: data, type: data type identifier}...], output: [{0: data, data type: data type identifier}, {1: data, data type: data type identifier}...]. The data type identifier is used to identify the type of data.

[0044] The image reading model is encapsulated as follows: Description: used to read images; input: [0: image, type: any of png, jpg, jpeg, bmp], output: [{0: image, type: IMAGE}].

[0045] The detection model is encapsulated as: Description: Detect targets in the input image. Supported detection targets include insulators, crossarms, tension clamps, suspension clamps, and shock-absorbing hammers. Input: [{0: image, type: IMAGE}], Output: [{0: insulator, type: detection box set}, {1: crossarm, type: detection box set}, {2: tension clamp, type: detection box set}, {3: suspension clamp, type: detection box set}, {4: shock-absorbing hammer, type: detection box set}].

[0046] The CLIP model is encapsulated as: Description: Given an image, a detection box set, and two statements, the model will judge the targets in the detection boxes respectively, and divide the detection boxes into two sets according to the judgment results. Input: [{0: image, type: IMAGE}, {1: detection box, type: detection box set}, {2: judgment statement 1, type: TEXT}, {3: judgment statement 2, type: TEXT}], Output: [{0: results that meet judgment statement 1, type: detection box set}, {1: results that meet judgment statement 2, type: detection box set}].

[0047] The output model is encapsulated as: Description: Used to merge the contents of multiple input nodes, input: [{0: node 0, type: Any}, {1: node 1, type: Any}...{N: node N, type: Any}], output: [{0: merged output, type: collection}]. Among them, any represents any type, and a node represents a sub-model.

[0048] I will not go into details about the encapsulation of other models. Please refer to the examples here.

[0049] For example 1, see Figure 3 , the power operation and inspection function information is processed based on the large language model in the aforementioned S102 to obtain the first functional module information in the large language model that can collaboratively implement the function corresponding to the power operation and inspection function information, which is represented as a specific example:

[0050] The power operation and inspection function information is used to identify the corroded insulators and corroded tension clamps in the operation and inspection diagram. The power operation and inspection function information is processed by a large language model to obtain the first function module information, wherein the first function module information includes: the image reading model name, and the data input format corresponding to the image reading model name is [0: image, type: any one of png, jpg, jpeg, bmp], and the data output format corresponding to the image reading model name is [{0: image, type: IMAGE}], the detection model name, and the data input format corresponding to the detection model name is [{0: image, type: IMAGE}], and the data output format corresponding to the detection model name is [{0: insulator, type: detection frame set}, {1: crossarm, type: detection frame set}, {2: tension clamp, type: detection frame set}, {3: suspension clamp, type : detection box set}, {4: shock-absorbing hammer, type: detection box set}], CLIP model name, the data input format corresponding to the CLIP model name is [{0: image, type: IMAGE}, {1: detection box, type: detection box set}, {2: judgment statement 1, type: TEXT}, {3: judgment statement 2, type: TEXT}], the data output format corresponding to the CLIP model name is [{0: result that meets judgment statement 1, type: detection box set}, {1: result that meets judgment statement 2, type: detection box set}], output model name, the data input format corresponding to the output model name is: [{0: node 0, type: Any}, {1: node 1, type: Any}…{N: node N, type: Any}], the data output format corresponding to the output model name is [{0: merged output, type: set}].

[0051] S103: Determine a first algorithm flow for implementing a function corresponding to the power operation and inspection function information according to the first function module information.

[0052] Optionally, in the aforementioned S103, the first algorithm flow for determining, according to the first functional module information, a function corresponding to the power operation and inspection functional information includes the following steps S1031-S1032:

[0053] S1031. Determine a first directed connection relationship between different first functional modules in the at least one first functional module based on a data input format and a data output format of each first functional module in the first functional module information;

[0054] Optionally, the first directed connection relationship refers to a directional connection relationship between different functional modules.

[0055] Optionally, the determining of a first directed connection relationship between different first function modules in the at least one first function module based on a data input format and a data output format of each first function module in the first function module information includes S0001-S0002:

[0056] S0001. Acquire a name of a second function module whose data input format and data output format both contain an image type from the data input format and data output format of each first function module in the first function module information;

[0057] Following the above example 1, see Figure 3 , obtaining from the first functional module information the name of an image reading model in which both the data input format and the data output format include the image type.

[0058] S0002. Establish a directed connection relationship by pointing the second function module name to the third function module name whose data input format is the same as the data output format of the second function module, or establish a first directed connection relationship by pointing the second function module name to the third function module name whose data input format includes the data output format of the second function module as the first function module corresponding to the second function module name and the first function module corresponding to the third function module name, and so on, to obtain the first directed connection relationship between different first function modules in the at least one first function module.

[0059] Following the above example 1, see Figure 3 , the first directed connection relationship between different first functional modules in at least one first functional module includes: the image reading model points to the detection model; the detection model points to the CLIP model; the CLIP model points to the output model. Specifically, the data output format of the image reading model is [{0: image, type: IMAGE}], and the data input format of the detection model is [{0: image, type: IMAGE}]. The data output format of the image reading model is the same as the data input format of the detection model, so the directed connection relationship from the image reading model to the detection model is taken as one of the first directed connection relationships.

[0060] The data output format of the detection model is [{0: insulator, type: detection frame set}, {1: cross arm, type: detection frame set}, {2: tension clamp, type: detection frame set}, {3: suspension clamp, type: detection frame set}, {4: shock-absorbing hammer, type: detection frame set}]; while the data input format of the CLIP model is [{0: image, type: IMAGE}, {1: detection frame, type: detection frame set}, {2: judgment statement 1, type: TEXT}, {3: judgment statement 2, type: TEXT}],

[0061] Among them, insulators, crossarms, tension clamps, suspension clamps and shock-absorbing hammers can all be used as detection frames, and their types are all detection frame sets. Therefore, the data output format of the detection model is included in the data input format of the CLIP model {1: detection frame, type: detection frame set}, which is equivalent to the data input format of the CLIP model including the data output format of the detection model. Therefore, the directed connection relationship from the detection model to the CLIP model is taken as one of the first directed connection relationships.

[0062] The data output format of the CLIP model is [{0: results that meet judgment statement 1, type: detection box set}, {1: results that meet judgment statement 2, type: detection box set}]. The data input format of the output model is: [{0: node 0, type: Any}, {1: node 1, type: Any}...{N: node N, type: Any}], where Any means any type. Therefore, it can be seen that the data input format of the output model includes the data output format of the CLIP model. Therefore, the directed connection relationship from the CLIP model to the output model is taken as one of the first directed connection relationships.

[0063] S1032: Use a first directed connection relationship between different first functional modules in the at least one first functional module as a first algorithm flow for implementing a function corresponding to the power operation and inspection functional information.

[0064] Specifically, the at least one first functional module is connected according to the first directed connection relationship to obtain a first algorithm flow for implementing the function corresponding to the power operation and inspection functional information.

[0065] Following the above example 1, see Figure 3 , Figure 3 The flowchart in is the first algorithm flow. Figure 3 The CLIP model in is used to detect insulators.

[0066] Optionally, after determining the first algorithm flow for implementing the function corresponding to the power operation and inspection function information according to the first function module information in the aforementioned S103, the method further includes the following steps S01-S02:

[0067] S01, sending the first algorithm flow to the client, and receiving a second algorithm flow after the relevant personnel of the client feedback confirms or amends the first algorithm flow, wherein the second algorithm flow includes at least one second function module name and a second directed connection relationship between different second function modules in at least one second function module;

[0068] S02. If the second algorithm process is different from the first algorithm process, a first algorithm model is constructed based on the second algorithm process, wherein the first algorithm model is used to analyze the input power operation and inspection diagram to obtain the detection result of the function corresponding to the power operation and inspection function information.

[0069] Specifically, different sub-models are connected according to the second algorithm flow to obtain a first algorithm model, wherein the first algorithm model is used to identify and process the power line hardware in the input power operation and inspection diagram to obtain a detection result of completing the function corresponding to the power operation and inspection function information.

[0070] Optionally, in the aforementioned S02, the constructing of the first algorithm model based on the second algorithm flow includes the following steps S021-S022:

[0071] S021. Obtain at least one second functional module corresponding to the at least one second functional module name in the second algorithm flow;

[0072] S022. According to a second directed connection relationship between different second functional modules in at least one second functional module, connect the different second functional modules to obtain a first algorithm model.

[0073] Optionally, if the second algorithm flow is the same as the first algorithm flow, a second algorithm model is constructed based on the first algorithm flow.

[0074] Specifically, if the second algorithm flow is the same as the first algorithm flow, it means that the model corresponding to the first algorithm flow generated based on the large model can realize the function corresponding to the power operation and inspection function information.

[0075] Optionally, if the second algorithm flow is different from the first algorithm flow, the second algorithm flow is reversed into second functional module information;

[0076] Specifically, if the second algorithm flow is different from the first algorithm flow, it means that the first algorithm flow generated based on the large model needs to be modified, and the model corresponding to the modified second algorithm flow can realize the function corresponding to the power operation and inspection function information.

[0077] Following the above example 1, see Figure 4 , Figure 4 The flowchart in FIG. 1 shows the second algorithm flow.

[0078] The large language model is fine-tuned based on the second functional module information to optimize the large language model.

[0079] Optionally, the explanation of the second functional module information can refer to the aforementioned explanation of the first functional module information, which will not be repeated here.

[0080] Continuing with the above exemplary embodiment, based on the first algorithm model obtained by the second algorithm process, which can realize the function corresponding to the power operation and inspection function information, the first algorithm model is used to process the power operation and inspection diagram, and the detection results are a set of rusted insulators and a set of rusted tension clamps.

[0081] In summary, by obtaining the power operation and inspection function information; processing the power operation and inspection function information based on the large language model, obtaining the first function module information in the large language model that can collaboratively realize the function corresponding to the power operation and inspection function information, the first function module information includes the first function module name of at least one first function module, the data input format and the data output format of each first function module, and the at least one first function module is included in the large language model; according to the first function module information, the scheme of the first algorithm flow for realizing the function corresponding to the power operation and inspection function information is determined. The algorithm flow of the power operation and inspection function is automatically generated by the large language model, and the developer does not need to encode the code to generate the algorithm flow, which saves manpower cost and time cost and improves the efficiency of generating the algorithm flow. Specifically, the relevant technicians only need to use natural language to describe the specific content and input and output requirements of the power operation and inspection function requirements. The method can automatically construct the algorithm flow according to the relevant information of the sub-modules that have been encapsulated in the large model; then, the developer performs simple hyperparameter modification and testing according to the generated algorithm flow, and the development of the algorithm flow can be completed. The entire process of developing the algorithm flow is modularized and standardized, so it can reduce the development threshold and workload, and significantly improve the development efficiency.

[0082] In addition, the correct algorithm flow after modification is serialized to obtain the second functional module information, and the large language model is fine-tuned based on the second functional module information. Through fine-tuning, the algorithm flow generated by the large language model is more accurate, thereby steadily improving the availability of the results of the automatic generation of the algorithm flow. By encapsulating each sub-model, a reusable sub-model is achieved, which improves the utilization rate of the sub-model, improves the resource utilization rate, and avoids the generation of more redundant resources.

[0083] Figure 5 A flowchart of another algorithm flow method provided for an exemplary embodiment of the present application, the execution subject of the method may be a client, and the method at least includes the following steps S301-S305:

[0084] S301, receiving a first algorithm flow sent by a server;

[0085] S302, relevant technical personnel verify whether the first algorithm process can realize the function corresponding to the power operation and inspection function information;

[0086] S303: If yes, sending the first algorithm flow as the second algorithm flow to the server;

[0087] S304: If not, obtaining information to be processed based on the first algorithm process and the power operation and inspection function information, wherein the information to be processed includes adding or deleting at least one third functional module, and adding or deleting a third directed connection relationship corresponding to the at least one third functional module;

[0088] Optionally, the newly added third functional module can modify hyperparameters, and relevant content about the hyperparameters can be found in the prior art.

[0089] S305: Based on the information to be processed and the first algorithm process, obtain the second algorithm process, and send the second algorithm process to the server.

[0090] For example, see Figure 4 When the relevant technical personnel verified that the first algorithm flow could not realize the function corresponding to the power operation and inspection function information, they directly added a new CLIP module in the first algorithm flow and modified the hyperparameters of the CLIP module to detect the tension clamp, and added a corresponding third directed connection relationship as a directed connection relationship between the detection model pointing to the CLIP model and the CLIP model pointing to the output model.

[0091] In summary, relevant technical personnel can add new first function modules and delete the original first function modules by themselves. The first algorithm process can be modified by modifying the hyperparameters. The modification process is carried out on the first algorithm process. Relevant technical personnel do not need to directly write code to complete the editing and debugging of the flowchart, which can significantly reduce the workload of developers, reduce the development threshold and repeated redundant code writing, and improve development efficiency. The algorithm model can be built by sending the flowchart tested by relevant technical personnel to the server, and the algorithm model can be put into production to realize the functions corresponding to the power operation and inspection function information. The efficiency and accuracy of generating algorithm processes are improved, and a lot of manpower and time costs are saved.

[0092] Figure 6 A structural diagram of an algorithm flow device provided for an exemplary embodiment of the present application; applicable to a server, wherein the device includes:

[0093] An acquisition unit 21 is used to acquire power operation and inspection function information;

[0094] The processing unit 22 is used to process the electric power operation and inspection function information based on the large language model to obtain first function module information in the large language model that can collaboratively implement the function corresponding to the electric power operation and inspection function information, wherein the first function module information includes a first function module name of at least one first function module, a data input format and a data output format of each first function module, and the at least one first function module is included in the large language model;

[0095] The determination unit 23 is used to determine a first algorithm flow for implementing the function corresponding to the power operation and inspection function information according to the first function module information.

[0096] Optionally, the device is used to determine the first algorithm flow for implementing the function corresponding to the power operation and inspection function information based on the first function module information, specifically for: determining the first directed connection relationship between different first function modules in the at least one first function module based on the data input format and data output format of each first function module in the first function module information; using the first directed connection relationship between different first function modules in the at least one first function module as the first algorithm flow for implementing the function corresponding to the power operation and inspection function information.

[0097] Optionally, after determining the first algorithm flow for implementing the function corresponding to the power operation and inspection function information according to the first function module information, the device is also used to: send the first algorithm flow to the client, and receive a second algorithm flow after confirmation or revision of the first algorithm flow by relevant personnel of the client for feedback, wherein the second algorithm flow includes at least one second function module name and a second directed connection relationship between different second function modules in at least one second function module; if the second algorithm flow is different from the first algorithm flow, constructing a first algorithm model based on the second algorithm flow, wherein the first algorithm model is used to analyze the input power operation and inspection diagram to obtain the detection result of the function corresponding to the power operation and inspection function information.

[0098] Optionally, the device is used to construct a first algorithm model based on the second algorithm process, specifically to: obtain at least one second functional module corresponding to the name of at least one second functional module in the second algorithm process; and connect the different second functional modules according to a second directed connection relationship between different second functional modules in at least one second functional module to obtain the first algorithm model.

[0099] Optionally, the device is further used to construct a second algorithm model based on the first algorithm process if the second algorithm process is the same as the first algorithm process.

[0100] Optionally, the device is also used to, if the second algorithm flow is different from the first algorithm flow, reverse the second algorithm flow into second functional module information; and fine-tune the large language model based on the second functional module information to optimize the large language model.

[0101] Figure 7 A schematic diagram of the structure of another algorithm flow device provided for an exemplary embodiment of the present application; applicable to a client, wherein the device includes:

[0102] A receiving unit 31 is used to receive a first algorithm flow sent by a server;

[0103] A verification unit 32, used by relevant technical personnel to verify whether the first algorithm process can realize the function corresponding to the power operation and inspection function information;

[0104] A sending unit 33 is configured to send the first algorithm flow as a second algorithm flow to a server if yes;

[0105] an acquisition unit 34, configured to, if not, acquire information to be processed based on the first algorithm process and the power operation and inspection function information, wherein the information to be processed includes adding or deleting at least one third functional module, and adding or deleting a third directed connection relationship corresponding to the at least one third functional module;

[0106] The acquisition unit 34 is further configured to acquire the second algorithm flow based on the information to be processed and the first algorithm flow, and send the second algorithm flow to the server.

[0107] It should be understood that the device embodiment and the method embodiment may correspond to each other, and similar descriptions may refer to the method embodiment. To avoid repetition, no further description is given here. Specifically, the device may perform the above method embodiment, and the above and other operations and / or functions of each module in the device are the corresponding processes in each method in the above method embodiment, respectively, and no further description is given here for the sake of brevity.

[0108] The above describes the device of the embodiment of the present application from the perspective of the functional module in conjunction with the accompanying drawings. It should be understood that the functional module can be implemented in hardware form, can be implemented by instructions in software form, and can also be implemented by a combination of hardware and software modules. Specifically, the steps of the method embodiment in the embodiment of the present application can be completed by the hardware integrated logic circuit and / or software form instructions in the processor, and the steps of the method disclosed in the embodiment of the present application can be directly embodied as a hardware decoding processor to perform, or a combination of hardware and software modules in the decoding processor to perform. Optionally, the software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, an electrically erasable programmable memory, a register, etc. The storage medium is located in a memory, and the processor reads the information in the memory, and completes the steps in the above method embodiment in conjunction with its hardware.

[0109] Figure 8 is a schematic block diagram of an electronic device provided in an embodiment of the present application, and the electronic device may include:

[0110] The memory 301 and the processor 302, the memory 301 is used to store the computer program and transmit the program code to the processor 302. In other words, the processor 302 can call and run the computer program from the memory 301 to implement the method in the embodiment of the present application.

[0111] For example, the processor 302 may be configured to execute the above method embodiments according to instructions in the computer program.

[0112] In some embodiments of the present application, the processor 302 may include but is not limited to:

[0113] General-purpose processor, digital signal processor (DSP), application-specific integrated circuit (ASIC), field programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic device, discrete hardware components, etc.

[0114] In some embodiments of the present application, the memory 301 includes but is not limited to:

[0115] Volatile memory and / or non-volatile memory. Among them, the non-volatile memory can be read-only memory (ROM), programmable ROM (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM) or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link DRAM (SLDRAM), and direct RAM bus random access memory (DR RAM).

[0116] In some embodiments of the present application, the computer program may be divided into one or more modules, which are stored in the memory 301 and executed by the processor 302 to complete the method provided by the present application. The one or more modules may be a series of computer program instruction segments capable of completing specific functions, and the instruction segments are used to describe the execution process of the computer program in the electronic device.

[0117] like Figure 8 As shown, the electronic device may also include:

[0118] The transceiver 303 may be connected to the processor 302 or the memory 301 .

[0119] The processor 302 may control the transceiver 303 to communicate with other devices, specifically, to send information or data to other devices, or to receive information or data sent by other devices. The transceiver 303 may include a transmitter and a receiver. The transceiver 303 may further include an antenna, and the number of antennas may be one or more.

[0120] It should be understood that the various components in the electronic device are connected via a bus system, wherein the bus system includes not only a data bus but also a power bus, a control bus and a status signal bus.

[0121] The present application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a computer, enables the computer to perform the method of the above method embodiment. In other words, the present application embodiment also provides a computer program product containing instructions, which, when executed by a computer, enables the computer to perform the method of the above method embodiment.

[0122] When software is used for implementation, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function according to the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (digital subscriber line, DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, server or data center. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more available media integration. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a digital video disc (digital video disc, DVD)), or a semiconductor medium (e.g., a solid state drive (solid state disk, SSD)), etc.

[0123] Those of ordinary skill in the art will appreciate that the modules 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.

[0124] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the module is only a logical function division. There may be other division methods in actual implementation, such as multiple modules 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 modules, which can be electrical, mechanical or other forms.

[0125] The modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical modules, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. For example, each first functional module in each embodiment of the present application may be integrated into a processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0126] The above are only specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. An algorithm flow method, characterized in that: For servers, including: Obtain power operation and inspection function information; Processing the electric power operation and inspection function information based on the large language model to obtain first function module information in the large language model that can collaboratively implement the function corresponding to the electric power operation and inspection function information, wherein the first function module information includes a first function module name of at least one first function module, a data input format and a data output format of each first function module, and the at least one first function module is included in the large language model; A first algorithm flow for implementing the function corresponding to the power operation and inspection function information is determined according to the first function module information.

2. The method according to claim 1, characterized in that The first algorithm process for determining, according to the first functional module information, a function corresponding to the power operation and inspection functional information, comprises: Determine a first directed connection relationship between different first functional modules in the at least one first functional module based on a data input format and a data output format of each first functional module in the first functional module information; The first directed connection relationship between different first functional modules in the at least one first functional module is used as a first algorithm flow for implementing the function corresponding to the power operation and inspection functional information.

3. The method according to claim 1, characterized in that After determining the first algorithm flow for implementing the function corresponding to the power operation and inspection function information according to the first function module information, the method further includes: The first algorithm flow is sent to the client, and the relevant personnel who receive the feedback from the client confirm the first algorithm flow or make a revised second algorithm flow, wherein the second algorithm flow includes at least one second function module name and a second directed connection relationship between different second function modules in at least one second function module; If the second algorithm flow is different from the first algorithm flow, a first algorithm model is constructed based on the second algorithm flow, wherein the first algorithm model is used to analyze the input power operation and inspection diagram to obtain the detection result of the function corresponding to the power operation and inspection function information.

4. The method according to claim 3, characterized in that The constructing a first algorithm model based on the second algorithm flow includes: Obtain at least one second functional module corresponding to the at least one second functional module name in the second algorithm flow; According to the second directed connection relationship between different second functional modules in at least one second functional module, the different second functional modules are connected to obtain a first algorithm model.

5. The method according to claim 3, characterized in that: If the second algorithm flow is the same as the first algorithm flow, a second algorithm model is constructed based on the first algorithm flow.

6. The method according to claim 3, characterized in that If the second algorithm flow is different from the first algorithm flow, the second algorithm flow is reversed into second functional module information; The large language model is fine-tuned based on the second functional module information to optimize the large language model.

7. An algorithm flow method, characterized in that, Applicable to clients, including: Receiving a first algorithm flow sent by a server; Relevant technical personnel verify whether the first algorithm process can realize the function corresponding to the power operation and inspection function information; If yes, sending the first algorithm flow as the second algorithm flow to the server; If not, obtaining information to be processed based on the first algorithm process and the power operation and inspection function information, wherein the information to be processed includes adding or deleting at least one third functional module, and adding or deleting a third directed connection relationship corresponding to the at least one third functional module; Based on the information to be processed and the first algorithm flow, the second algorithm flow is obtained, and the second algorithm flow is sent to the server.

8. An algorithm flow device, characterized in that: For servers, including: An acquisition unit, used for acquiring power operation and inspection function information; a processing unit, configured to process the electric power operation and inspection function information based on a large language model, and obtain first function module information in the large language model that can collaboratively implement a function corresponding to the electric power operation and inspection function information, wherein the first function module information includes a first function module name of at least one first function module, a data input format and a data output format of each first function module, and the at least one first function module is included in the large language model; A determination unit is used to determine a first algorithm flow for implementing the function corresponding to the power operation and inspection function information according to the first function module information.

9. An electronic device, characterized in that: include: processor; as well as A memory, configured to store executable instructions of the processor; The processor is configured to perform the method of any one of claims 1 to 7 by executing the executable instructions.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images