Method and device for generating business execution component through computing power of intelligent computing center

Abstract

The invention provides a method and device for generating a service execution component through computing power of an intelligent computing center, and relates to the technical field of intelligent computing centers, intelligent computing centers and computing power infrastructure, the method comprises the following steps: S1, based on the computing power of the intelligent computing center, obtaining coding demand information from static planning information, the static planning information is used for indicating a business processing logic of a target business; and S2, taking the coding demand information as a coding prompt word, and carrying out coding based on the computing power of an intelligent computing center to obtain an adapter function. Under the condition that the basic function cannot meet all the requirements of the target service, the personalized function requirements of the target service are obtained from the static planning information corresponding to the target service by using the computing power of the intelligent computing center to form the coding requirement information, and the coding requirement information is used as the coding prompt word, so that the coding efficiency is improved. And quickly coding to obtain an adapter function for connecting the basic function, so as to guarantee smooth execution of static planning.

Description

Technical Field

[0001] The present invention relates to the technical fields of intelligent computing centers, intelligent computing centers, and computing power infrastructure, and in particular, to a method and apparatus for generating service execution components through the computing power of an intelligent computing center. Background Art

[0002] With the rapid development of artificial intelligence technology, "intelligent computing centers" and "intelligent computing centers" have emerged as the times require.

[0003] An "intelligent computing center" refers to a facility that uses large-scale heterogeneous computing power resources, including general computing power and intelligent computing power, and mainly provides the required computing power, data, and algorithms for artificial intelligence applications (such as scenarios of artificial intelligence deep learning model development, model training, and model inference, etc.). The intelligent computing center covers facilities, hardware, and software, and can provide full-stack capabilities from underlying computing power to top-level application enablement.

[0004] The "intelligent computing center" includes, but is not limited to, the "intelligent computing center".

[0005] An "intelligent computing center", that is, an artificial intelligence computing center, is a type of computing power infrastructure that is based on artificial intelligence theory, adopts an artificial intelligence computing architecture, and provides computing power services, data services, and algorithm services required for artificial intelligence applications.

[0006] "Computing power" is the core of "intelligent computing centers" and "intelligent computing centers". It is the ability of computer devices or computing / data centers to process information, the ability of computer hardware and software to cooperate to jointly execute a certain computing requirement, the computing ability to achieve the output of target results through processing information data, and a new type of productive force that integrates information computing power, network carrying capacity, and data storage capacity. It mainly provides services to society through computing power infrastructure.

[0007] Currently, how to quickly generate an agent that matches the requirements is an urgent problem to be solved. A possible solution is as follows: By means of custom coding, a batch of general functions and the corresponding function patterns are pre-constructed. This part of the pre-constructed functions can be called base functions, and the corresponding function patterns can be called base function schemas. By providing the base function schemas to the large model, the large model can master the uses and usages of the base functions. As the business processing core of the pre-generated agent, the large model can, after the agent receives the business requirements input by the user, dynamically form a static plan that meets the above business requirements based on the mastered uses and usages of the base functions, and execute the static plan to achieve a dynamic response to the business requirements. In actual scenarios, the base functions may not cover all business requirements. To address this issue, the large model can temporarily write code to generate adapter functions that can fill the gaps in the base functions, so as to cooperate with the base functions to achieve comprehensive coverage of the business requirements.

[0008] In applications, it is found that when the prior art uses a large model for function coding work, not only is the coding efficiency and coding accuracy low, but also the cooperation with existing functions is usually not considered. Therefore, if the prior art is used for coding adapter functions, the adapter functions may not be able to effectively connect the output of the previous step and / or the input of the subsequent step, which will result in the automatic execution of the static plan not being smooth.

[0009] It can be seen that there are problems in the prior art such as low coding efficiency of adapter functions, poor coding effect, and poor data connection effect with the functions corresponding to adjacent steps. Summary of the Invention

[0010] The purpose of the present invention is to provide a method and device for generating business execution components through the computing power of an intelligent computing center, which are used to solve the problems in the prior art such as low coding efficiency of adapter functions, poor coding effect, and poor data connection effect with the functions corresponding to adjacent steps.

[0011] To solve the above technical problems, the present invention is implemented as follows:

[0012] In a first aspect, the present invention provides a method for generating business execution components based on the computing power of an intelligent computing center, including:

[0013] Step S1: Obtain the encoding requirement information from the static planning information based on the computing power of the intelligent computing center. The static planning information is used to indicate the business processing logic of the target business. The encoding requirement information includes: the function name of the adapter function to be encoded, the step description text of the step corresponding to the adapter function to be encoded in the static planning information, the function output path of the adapter function to be encoded, the function output mode of the adapter function to be encoded, and the function input information of the adapter function to be encoded. The target business includes a vision processing business based on a computer vision agent, and the encoding requirement information is used to represent the personalized function requirements in the vision processing business.

[0014] Step S2: Use the encoding requirement information as an encoding prompt word and encode the adapter function based on the computing power of the intelligent computing center. The adapter function is used to cooperate with a preset basic function to execute the target business. The basic function is used to implement the general functions in the vision processing business, and the adapter function is used to implement the personalized functions in the vision processing business.

[0015] In one embodiment, the function output mode of the adapter function to be encoded does not include the parameter explanation text of the output parameters of the adapter function to be encoded.

[0016] The function output mode of the adapter function to be encoded includes at least one of the following:

[0017] The parameter name of the output parameter of the adapter function to be encoded;

[0018] The data format of the output parameter of the adapter function to be encoded;

[0019] The value range of the output parameter of the adapter function to be encoded.

[0020] In one embodiment, the function input information of the adapter function to be encoded includes: the step output of the previous step associated with the adapter function, and / or, the business data to be processed carried in the business requirement information corresponding to the target business, where the execution order of the previous step is earlier than the execution order of the step corresponding to the adapter function.

[0021] In one embodiment, the encoding requirement information further includes the function code template of the adapter function to be encoded.

[0022] The function code template of the adapter function to be encoded sequentially includes:

[0023] The head identifier indicating the first digit of the adapter function to be encoded;

[0024] The import package code text indicating the function dependencies of the adapter function to be encoded;

[0025] The parameter code text of the function input that indicates the adapter function to be encoded;

[0026] The annotation code text of the function annotation that indicates the adapter function to be encoded;

[0027] The function body code text of the function function that indicates the adapter function to be encoded;

[0028] The parameter code text of the function output that indicates the adapter function to be encoded;

[0029] The tail identifier indicating the end of the adapter function to be encoded.

[0030] In one embodiment, the encoding requirement information further includes: a plurality of input parameter names and a plurality of output parameter names;

[0031] Wherein, the plurality of input parameter names are used to indicate: a plurality of input parameters corresponding to the adapter function in the static planning information;

[0032] The plurality of output parameter names are used to indicate: a plurality of output parameters corresponding to the adapter function in the static planning information.

[0033] In one embodiment, the step S2 includes:

[0034] Step S21, using the encoding requirement information as an encoding prompt word, iteratively execute the function encoding task based on the computing power of the intelligent computing center;

[0035] The nth iteration in the iteration includes:

[0036] Step S211, when the function encoding task in the nth iteration is successfully executed, form the adapter function based on the execution result of the function encoding task in the nth iteration, and exit the iteration;

[0037] Step S212, when the function encoding task in the nth iteration fails and n is less than or equal to the set iteration times threshold, use the task failure information of the function encoding task in the nth iteration and the encoding requirement information as encoding prompt words, and use the computing power of the intelligent computing center to execute the function encoding task in the (n + 1)th iteration;

[0038] Step S213, when the function encoding task in the nth iteration fails and n is greater than the iteration times threshold, exit the iteration, and output the task failure information of the function encoding task in the nth iteration;

[0039] Wherein, n is an integer greater than or equal to 1, and when n = 1, the function encoding task in the nth iteration uses the encoding requirement information as an encoding prompt word.

[0040] In a second aspect, the present invention further provides a device for a computing power generation service execution component based on an intelligent computing center, including:

[0041] A requirement acquisition module, configured to obtain encoded requirement information from static planning information based on the computing power of the intelligent computing center, where the static planning information is used to indicate the business processing logic of the target service, and the encoded requirement information includes: the function name of the adapter function to be encoded, the step description text of the corresponding step of the adapter function to be encoded in the static planning information, the function output path of the adapter function to be encoded, the function output mode of the adapter function to be encoded, the function input information of the adapter function to be encoded, the target service includes a vision processing service based on a computer vision agent, and the encoded requirement information is used to represent the personalized function requirements in the vision processing service;

[0042] An encoding module, configured to use the encoded requirement information as an encoding prompt word and encode an adapter function based on the computing power of the intelligent computing center, where the adapter function is used to cooperate with a preset basic function to execute the target service, the basic function is used to implement the general functions in the vision processing service, and the adapter function is used to implement the personalized functions in the vision processing service.

[0043] In a third aspect, the present invention further provides a server, including: a processor, a memory, and a program stored on the memory and executable on the processor, and when the program is executed by the processor, it implements the steps in the method for generating a service execution component based on the computing power of the intelligent computing center as described in the first aspect above.

[0044] In a fourth aspect, the present invention further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps in the method for generating a service execution component based on the computing power of the intelligent computing center as described in the first aspect above.

[0045] In a fifth aspect, the present invention provides a computer program product, including computer instructions, and when the computer instructions are executed by a processor, it implements the steps in the method for generating a service execution component based on the computing power of the intelligent computing center as described in the first aspect above.

[0046] In the present invention, when the basic function cannot meet all the requirements of the target service, the computing power of the intelligent computing center is utilized to obtain the personalized function requirements of the target service from the static planning information corresponding to the target service to form coding requirement information, and the coding requirement information is used as a coding prompt word to quickly code an adapter function for supplementing the function of the basic function. Among them, by defining that the coding requirements include the function name of the adapter function to be coded, the step description text of the corresponding step, the function output path, the function output mode, and the function input information, accurate and brief prompts for the static planning and the actual execution results of the previous workflow can be realized, so as to improve the usability of the obtained adapter function, and significantly improve the coding efficiency, effect of the adapter function, and the data connection effect between the functions corresponding to adjacent steps. Furthermore, the fluency of automatically executing the static planning is also improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] By reading the following detailed description of the preferred embodiments, various other advantages and benefits will become clear to those of ordinary skill in the art. The drawings are only for the purpose of showing the preferred embodiments and are not considered to be a limitation of the present invention. Moreover, throughout the drawings, the same reference numerals are used to represent the same components. In the drawings:

[0048] Figure 1 is an architecture diagram of a two-stage intelligent agent automation workflow provided by the present invention;

[0049] Figure 2 is a schematic diagram of the results of an experimental intelligent agent automation workflow using Qwen2.5-72B-Instruct provided by the present invention;

[0050] Figure 3 is a schematic flowchart of a method for generating service execution components by the computing power of an intelligent computing center provided by the present invention;

[0051] Figure 4 is a schematic structural diagram of a device for generating service execution components by the computing power of an intelligent computing center provided by the present invention;

[0052] Figure 5 is a schematic structural diagram of an electronic device provided by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0053] The technical solutions in the present invention will be clearly and completely described below with reference to the drawings in the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art without making creative efforts based on the embodiments in the present invention belong to the scope of protection of the present invention.

[0054] First, the technical terms related to the present invention will be briefly described below.

[0055] The "computing power" referred to in the present invention is the ability of a computer device or a computing / data center to process information. It is the ability of computer hardware and software to cooperate to jointly execute a certain computing requirement. It is the computing ability to process information data and achieve the output of the target result. It is a new type of productive force integrating information computing power, network carrying capacity, and data storage capacity, and mainly provides services to society through the computing power infrastructure.

[0056] The "computational power" (Computational Power, CP) referred to in the present invention is an ability of the data center server to process data and achieve result output. It is a comprehensive index to measure the computing ability of the data center, including general computing ability, supercomputing ability, and intelligent computing ability. The commonly used measurement unit is the number of floating-point operations per second (FLOPS: Floating Point Operations Per Second, 1EFLOPS = 10^18 FLOPS). The larger the value, the stronger the comprehensive computing ability. It is estimated that 1EFLOPS is approximately the computing power output of 5 Tianhe 2A or 500,000 mainstream server CPUs or 2 million mainstream laptops. The calculation formula is: CP = CP 通用 +CP 智能 +CP 超级 .

[0057] The "network power" (Network Power, NP) referred to in the present invention is the manifestation of the data transmission ability of the computing power facility, including the comprehensive ability such as network architecture, network bandwidth, transmission delay, intelligent management and scheduling, etc. The network power involves the network transmission inside and between data centers and is a comprehensive index to measure the network transmission scheduling ability. In the present invention, the network power uses the video memory bandwidth.

[0058] The "storage power" (Storage Power, SP) referred to in the present invention is the comprehensive ability of the data center in four aspects: data storage capacity, performance, security and reliability, and green and low-carbon. It is a comprehensive index to measure the data storage ability of the data center, including external storage devices such as storage arrays and server internal storage devices. The commonly used measurement unit for storage capacity is exabyte (EB, 1EB = 2^60 bytes). The commonly used measurement unit for performance is the number of read and write operations per second per unit capacity (Input / Output Operations Per Second / TB, IOPS / TB). The disaster recovery ratio is an important manifestation of security and reliability.

[0059] The "computing power infrastructure" described in the present invention is a new type of information infrastructure that integrates information computing power, network carrying capacity, and data storage capacity, and can realize the centralized computing, storage, transmission, and application of information.

[0060] The "new type of information infrastructure" described in the present invention refers to: mainly including network infrastructures such as 5G networks, fiber broadband networks, backbone networks, international communication networks, satellite Internet, etc., computing power infrastructures such as data centers, general computing power centers, intelligent computing centers, supercomputing centers, etc., and new technology infrastructures such as artificial intelligence, blockchain, quantum computing, etc.

[0061] The "computing power" described in the present invention includes general computing power, intelligent computing power, and super computing power.

[0062] The "general computing power" described in the present invention is the computing power provided by servers based on central processing unit (CPU) chips, and is used to support basic general computing such as cloud computing and edge computing.

[0063] The "intelligent computing power" described in the present invention is for various artificial intelligence innovation applications, and is a computing platform deployed on a large scale based on dedicated chips such as graphics processing unit (GPU), field programmable gate array (FPGA), and application specific integrated circuit (ASIC). For example, natural language processing, machine vision, etc.

[0064] The "super computing power" described in the present invention is mainly the computing power provided by high-performance computing clusters such as supercomputers. It utilizes the centralized computing resources of multiple computer systems working in parallel and processes extremely complex or data-intensive problems through a dedicated operating system. It is mainly used for computing in cutting-edge scientific fields, such as planetary simulation, drug molecule design, gene analysis, etc.

[0065] The "intelligent computing center" described in the present invention refers to a facility that provides the required computing power, data, and algorithms mainly for artificial intelligence applications (such as scenarios of artificial intelligence deep learning model development, model training, and model inference) by using large-scale heterogeneous computing power resources, including general computing power (CPU) and intelligent computing power (GPU, FPGA, etc.). The intelligent computing center covers facilities, hardware, and software, and can provide full-stack capabilities from underlying computing power to top-level application enabling.

[0066] The "intelligent computing center" described in the present invention includes but is not limited to the "intelligent computing center".

[0067] The "Intelligent Computing Center" described in the present invention, namely the artificial intelligence computing center, is a type of computing infrastructure based on artificial intelligence theory, adopting an artificial intelligence computing architecture, and providing computing power services, data services, and algorithm services required for artificial intelligence applications.

[0068] The "Computing Power Center" described in the present invention refers to: a facility mainly composed of infrastructure such as wind, fire, water, and electricity and IT software and hardware devices, with computing power, carrying capacity, and storage capacity, including general data centers, intelligent computing centers, supercomputing centers, etc.

[0069] The "Supercomputing Center" described in the present invention refers to: namely the supercomputing data center, which is a data center based on supercomputers or large-scale computing clusters, capable of providing functions such as large-scale computing, storage, and network services, and is widely used in application scenarios such as aerospace, national defense, oil exploration, climate modeling, and genome sequencing.

[0070] The "Computing Power Resources" described in the present invention refer to the technologies and facilities required for the development of the digital society with information computing, transmission, storage, and application capabilities, including but not limited to computing resources such as CPUs and GPUs, network resources such as switches and routers, storage resources such as storage arrays and distributed storage, security resources such as firewalls and intrusion detection systems, and support and guarantee resources such as wind, fire, water, and electricity.

[0071] The "Models" and "Large Models" described in the present invention include but are not limited to "Large Language Models" and "Multimodal Large Models".

[0072] The "Large Language Model" described in the present invention refers to the large language model (LLM), which is a language model with a relatively large number of parameters, aiming to understand and generate human language, trained with a large amount of text data, and can perform a wide range of tasks including text summarization, translation, sentiment analysis, etc.

[0073] The "Multimodal Large Models" (multimodal large models) described in the present invention refer to: models that jointly train multi-modal information such as text, images, videos, and audio, including but not limited to multimodal large language models.

[0074] The "agent" described in the present invention refers to an entity that can perceive the environment and take actions to achieve specific goals. It can be software, hardware, or a system, and has autonomy, adaptability, and interaction capabilities. The agent perceives changes in the environment (such as through sensors or data input), makes judgments and decisions based on the knowledge and algorithms it has learned, and then executes actions to affect the environment or achieve a predetermined goal. Agents are widely used in the field of artificial intelligence, commonly found in automation systems, robots, virtual assistants, and game characters, etc. The core lies in the ability to learn autonomously and continuously evolve to better complete tasks and adapt to complex environments.

[0075] The "base function" described in the present invention refers to a function pre-constructed by developers according to the common / high-frequency business operations of the business scenarios faced by the agent. For example, when the agent is specifically a computer vision agent, the common / high-frequency business operations of the business scenarios faced by the agent may include: image preprocessing operations, image segmentation operations, image noise filtering operations, image target detection operations, etc.

[0076] The "base function schema" described in the present invention refers to the "base function usage manual" formed by developers based on the functions and usage of the base functions. The base function schema can include items such as name, description, type, etc. Among them, the name item can include the function name, the input parameter names of the function, the output parameter names of the function, etc. The description item can be understood as the function description, the semantic description of the input parameters of the function, and the semantic description of the output parameters of the function; the type item includes the data format description of the input parameters of the function and the data format description of the output parameters of the function. The large model can master the uses and usages of the base functions through the "base function schema".

[0077] The "planner" described in the present invention refers to a large model for task planning. After the task rule large model masters the uses and usages of the base functions through the base function schema, it can be used to generate a static plan in one go. The static plan contains all the business processing steps to complete the business requirements, and each business processing step contains its explanation (also known as the step description text), the method name called, and the input parameter information (each input parameter and its corresponding assignment).

[0078] The "autocoder" described in the present invention refers to a large model for code writing, which is used to write adapter functions.

[0079] The "business execution component" described in the present invention refers to an adapter function encoded by an autoencoder according to encoding requirement information, and this adapter function is used to cooperate with a basic function to achieve comprehensive coverage of business requirements. The above-mentioned encoding requirement information includes: the business processing requirements of the business processing operations to be completed by the adapter function, and the data transmission requirements of the adjacent business processing steps in the static plan for the adapter function. Therefore, the adapter function can effectively connect adjacent business processing steps in the static plan to ensure the smooth execution of the static plan;

[0080] It should be noted that the adjacent business processing steps referred to in the present invention should be understood as: the pre-order steps associated with the adapter function in the static plan, and / or, the post-order steps associated with the adapter function in the static plan. Among them, the pre-order steps associated with the adapter function in the static plan are specifically the pre-order steps whose step outputs are used as the function inputs of the adapter function. Similarly, the post-order steps associated with the adapter function in the static plan are specifically: the post-order steps that use the function output of the adapter function as the step input; among them, the pre-order steps refer to any business processing steps in the static plan that are located before the adapter function, and the post-order steps refer to any business processing steps in the static plan that are located after the adapter function.

[0081] The "static planning information" described in the present invention refers to the data representation of the static plan, which at least includes a plurality of step data bodies arranged in an orderly manner. The plurality of step data bodies correspond one-to-one to a plurality of business processing steps, and the plurality of business processing steps are executed in sequence to complete the user's business requirements.

[0082] The "step data body" described in the present invention is the data representation of a corresponding business processing step, including but not limited to: the step description text of the corresponding business processing step (used to describe the business processing step to indicate the function, purpose, and / or execution method of the step, so as to facilitate users or developers to understand and implement), the step number of the corresponding business processing step among the plurality of business processing steps, the input data of the corresponding business processing step (such as the data name, data type requirements, data source, etc. of the input data), and the data processing method of the corresponding business processing step (calling a specific method through the method name).

[0083] The "pipeline" in the present invention refers to the process of executing the static plan.

[0084] Regarding the problem of quickly generating an agent that matches user needs, the present invention provides an agent automation workflow solution, which specifically includes: By means of custom coding, a batch of general functions and the corresponding function patterns are pre-constructed in advance. This part of the pre-constructed functions can be called base functions, and the corresponding function patterns of the functions can be called base function schemas. By providing the base function schemas to the large model, the large model can master the uses and usages of the base functions. As the business processing core of the pre-generated agent, the large model can, after the agent receives the business needs input by the user, dynamically form a static plan that meets the above-mentioned business needs based on the mastered uses and usages of the base functions, and execute the static plan to achieve a dynamic response to the business needs; In actual scenarios, the base functions may not cover all business needs. To address this problem, the large model can temporarily write code to generate adapter functions that can fill the gaps in the base functions to cooperate with the base functions to achieve comprehensive coverage of business needs.

[0085] In one example, the above-mentioned agent automation workflow solution is implemented by leveraging the powerful computing power and flexible resource allocation of the intelligent computing center. The corresponding architecture of the two-stage auto-pipeline can be as Figure 1 shown Figure 1 described in the relevant diagrams as follows:

[0086] One user query: Users input their requests or demand content.

[0087] Planner: Used to parse user needs and formulate an execution plan (plan). If the parsing fails, it may output an error that cannot be converted into JSON format.

[0088] Plan list: A list of specific steps generated according to user needs.

[0089] Retrieve all-step function dependencies: Retrieve the function dependencies related to each step to ensure that all required functions are available.

[0090] Get executable input values:

[0091] Given values: Values directly provided to the function.

[0092] User inputs: Dynamic input values provided by the user in the query.

[0093] Previous function outputs: Function output values from the previous step.

[0094] Media download: This step can also support the download of additional required media content.

[0095] Execute one step: Refers to the indication of the execution details of a specific step.

[0096] Execute function: Used to actually call the function to perform data processing work based on the obtained input values.

[0097] Base function: Performs specific basic functions and supports the agent to have the set basic business processing capabilities.

[0098] Adapter function (autoencoder): Serves as a supplement to the base function.

[0099] Retrieve depended function schemas: Obtain the structures or schemas of other dependent functions.

[0100] Autoencoder coding: Automatically generate the code involved in the above process.

[0101] Exception code: If an error or exception occurs, the system will return a specific exception code for subsequent analysis and debugging. For example:

[0102] Exceptions can be classified into the following 6 categories:

[0103] 0: No exception, the pipeline execution is successful

[0104] 1: Other pipeline exceptions

[0105] 2: Autoencoder execution failed

[0106] 3: Base function execution failed

[0107] 4: Plan acquisition failed (usually due to the failure of converting the plan to JSON)

[0108] 5: The placeholder expression of the variable in the plan is incorrect and the parsing fails.

[0109] It should be noted that applying the two-stage auto-pipeline has at least the following advantages:

[0110] 1. Higher accuracy: There are more cases where the given plan can be successfully executed;

[0111] 2. By decomposing tasks, the total inference time of the LLM can be shortened;

[0112] 3. By manually handling the core information transfer between steps, the scope of effective information can be clarified and the transparency of the system can be increased.

[0113] The results of experimenting with the two-stage auto-pipeline using the Qwen2.5-72B-Instruct tool for different data sources are as Figure 3 shown.

[0114] Experiments found that errors mainly occurred in the basic function execution stage and the variable placeholder expression stage. Among them, the errors in the basic function execution stage included: incorrect parameter passing, confusion in using the picture list and video path, and some parameters unrelated to the input (such as thresholds), etc.; the errors in the variable placeholder expression stage included: importing external packages (lacking package dependencies), incorrect understanding and processing of complex parameters, etc. In addition, some errors were caused by poor compliance with the json format of the plan output.

[0115] In the automatic coding link of the above solution, since the existing technology usually does not consider the cooperation between the encoding function and the existing functions, the existing technology cannot well support the implementation of the automatic coding link. More specifically, the encoding scheme provided by the existing technology has problems such as low encoding efficiency of the adapter function, poor encoding effect, and poor data connection effect between the functions corresponding to adjacent steps.

[0116] To improve the problems of low encoding efficiency of the adapter function, poor encoding effect, and poor data connection effect between the functions corresponding to adjacent steps, the present invention provides a method for generating a service execution component based on the computing power of an intelligent computing center, so as to, on the premise of utilizing the powerful computing power of the intelligent computing center, through corresponding encoding process planning and setting of hint schemes, guide the intelligent computing center to automatically complete the encoding work of the adapter function, which can not only improve the encoding efficiency and encoding effect of the adapter function, but also better realize the data connection between the adapter function and the functions corresponding to adjacent steps.

[0117] Please refer toFigure 3 , Figure 3 is a method for generating a computing power generation service execution component based on an intelligent computing center provided by the present invention. As Figure 3 shown, it includes the following steps:

[0118] Step S1: Obtain coding requirement information from static planning information based on the computing power of the intelligent computing center. Among them, the static planning information is used to indicate the business processing logic of the target business. The coding requirement information includes: the function name of the adapter function to be coded, the step description text of the step corresponding to the adapter function to be coded in the static planning information, the function output path of the adapter function to be coded, the function output mode of the adapter function to be coded, the function input information of the adapter function to be coded. The target business includes a visual processing business based on a computer vision agent, and the coding requirement information is used to represent the personalized function requirements in the visual processing business.

[0119] The above target business is used to meet the business needs of users. For example, performing oil stain detection on the image / image set input by the user, performing face recognition and verification on the image / image set provided by the user, etc.

[0120] The personalized function requirements in the visual processing business can be understood as: the function requirements that cannot be realized by basic functions among all the function requirements corresponding to the target business.

[0121] The above static planning information is generated by a "task planning large model" (planner). The input of the planner is the business requirement information of the target business, and the output of the planner is a static plan indicating the business processing logic of the target business. The data representation of this static plan is the static planning information.

[0122] As mentioned above, the static planning information includes a plurality of step data bodies corresponding one by one to the multiple business processing steps forming the business processing logic of the target business. The action of obtaining coding requirement information from the static planning information can be further described as:

[0123] Among the multiple step data bodies, identify the step data bodies that cannot be realized by basic functions, and analyze the identified step data bodies and other step data bodies associated with this step data body to determine the requirement data such as the step input, step output, function name of the function used in the step, and function processing requirements of the function used in the step, and accordingly form coding requirement information.

[0124] Among them, the encoding requirement information includes the function name of the adapter function to be encoded, which is used to limit the naming of the adapter function to ensure that the generated adapter function can be correctly called corresponding to the steps in the static planning information, thereby ensuring that the adapter function can be correctly executed and avoiding the problem of function call failure caused by function name differences.

[0125] The encoding requirement information includes the step description text of the adapter function to be encoded corresponding to the steps in the static planning information, which is used to limit the function of the adapter function to ensure that the function of the generated adapter function can meet the function requirements of the corresponding steps in the static planning information, ensure that the adapter function can output the expected results, and thus ensure the smooth execution of the static plan.

[0126] The encoding requirement information includes the function output path of the adapter function to be encoded to match possible output storage requirements (for example, when the corresponding step of the adapter function is the last step of multiple steps included in the target service), to avoid unnecessary errors caused by the lack of the function output path of the adapter function.

[0127] The encoding requirement information includes the function output mode of the adapter function to be encoded to standardize the function output format of the adapter function, so as to smoothly complete the data connection between the adapter function and the associated previous and / or subsequent steps. Compared with the method of restricting the function output format in text form, restricting the function output format in mode form is more accurate.

[0128] The encoding requirement information includes the function input information of the adapter function to be encoded to introduce a more concise and real function input, to avoid complex function input text interfering with the large model's understanding of the function input, and ensure the correct understanding and use of the function input by the large model.

[0129] It should be noted that in the application, when the function output of the adapter function to be encoded is referenced by the basic function in the subsequent steps and there are data type requirements at the reference position, the function output mode of the adapter function to be encoded is dynamically generated according to the parameter mode (schema) corresponding to the reference position; otherwise, the function output mode of the adapter function to be encoded is one of the pre-set one or more standard output modes.

[0130] Exemplarily, the visual processing service may be an image recognition service, a target detection service, an image segmentation service, a pose estimation service, a face recognition service, an image generation service, a video analysis service, an augmented reality service, an autonomous driving service, etc.

[0131] For example, when the target business is the oil pollution detection business for power companies, the corresponding requirement description text can be: The oil pollution detection agent produced by the power company needs to process the input images and detect whether there is oil pollution in 4 specific components (riser base, pressure relief valve, cooler, bushing) in the transformer and reactor. The task can be divided into the following steps: image preprocessing, object detection, image segmentation, oil pollution detection, output of detection results, and generation of alarms.

[0132] The static planning information generated according to this requirement description text can be:

[0133]

[0134]

[0135] The above static planning information includes 5 step data bodies arranged in order, corresponding to the 5 business processing steps of image preprocessing, object detection, image segmentation, oil pollution detection, output of detection results, and generation of alarms respectively. In the step data body corresponding to image preprocessing, the string "Preprocess the input images to ensure they are suitable for further processing." can be understood as the step description text, the string "preprocess" is the method name of the business processing method used, the character "<|input_step_1.image_path_list|>" indicates that this step requires user input, and the variable name of the input variable is "image_path_list", and the string "resize_dims":[800,600]} indicates the size requirement of the input image; the meanings of the step data bodies corresponding to steps such as object detection, segmentation, oil pollution detection, and output of detection results are similar. To avoid repetition, they will not be elaborated here.

[0136] In this example, if it is assumed that the 4th step, "Oil pollution detection", cannot be implemented by the basic function, then it is necessary to automatically code the corresponding adapter function for the 4th step, "Oil pollution detection". At this time, the function name of the adapter function to be coded is "postprocess"; the step description text of the step corresponding to the adapter function to be coded in the above static planning information is "Post-process the detection and segmentation results, screen the oil pollution areas, and classify them."; the function output path of the adapter function to be coded is output_step_4.predict_img_list; the function output mode of the adapter function to be coded is the mode corresponding to the parameter "image_path_list"; the function input information of the adapter function to be coded is the output result of the 3rd step, "Image segmentation" (obtained through the path output_step_3.predict_img_list).

[0137] Step S2: Using the encoding requirement information as an encoding prompt word, based on the computing power of the intelligent computing center, an adapter function is encoded, where the adapter function is used to cooperate with a preset basic function to execute the target service, the basic function is used to implement general functions in the vision processing service, and the adapter function is used to implement personalized functions in the vision processing service.

[0138] In the case where the basic function cannot meet all the requirements of the target service, the present invention utilizes the computing power of the intelligent computing center to obtain the personalized function requirements of the target service from the static planning information corresponding to the target service to form encoding requirement information, and uses this encoding requirement information as an encoding prompt word to quickly encode an adapter function for supplementing the functions of the basic function. Among them, by specifying that the encoding requirements include the function name of the adapter function to be encoded, the step description text of the corresponding step, the function output path, the function output mode, and the function input information, it is possible to accurately and briefly prompt the static planning and the actual execution results of the previous workflow, so as to improve the usability of the obtained adapter function, and significantly improve the encoding efficiency, effect of the adapter function, and the data connection effect between the functions corresponding to adjacent steps, thereby also improving the fluency of automatically executing the static planning.

[0139] In one embodiment, the function output mode of the adapter function to be encoded does not include the parameter explanation text of the output parameters of the adapter function to be encoded;

[0140] The function output mode of the adapter function to be encoded includes at least one of the following:

[0141] The parameter name of the output parameter of the adapter function to be encoded;

[0142] The data format of the output parameter of the adapter function to be encoded;

[0143] The value range of the output parameter of the adapter function to be encoded.

[0144] Among them, the parameter explanation text of the output parameter of the adapter function to be encoded is used to represent semantic information such as the purpose / business meaning of the output parameter of the adapter function to be encoded. For example, when the output parameter of the adapter function to be encoded refers to an image, the corresponding parameter explanation text can describe the image as "the road image to be recognized" or "the monitored image after binarization processing", etc.

[0145] In this embodiment, by setting the function output mode of the adapter function to be encoded not to include the parameter explanation text of the output parameter of the adapter function to be encoded, especially when the function output of the adapter function to be encoded is referenced by the base function in subsequent steps and there are data type requirements at the reference position, it is possible to effectively avoid the mixing of interference information unrelated to function encoding, so as to avoid the probability of confusion in the encoding process of the adapter function and improve the accuracy and reliability of the encoded adapter function.

[0146] The setting of the parameter name, data format, and value range in the function output mode can effectively standardize the function output format of the adapter function, so as to avoid problems such as the output parameters included in the function output of the adapter function not matching the requirements in the static specification information, such as the parameter name of the output parameter not matching, the data format of the output parameter not matching, and the specific value of the output parameter not matching the set value range; thereby ensuring the accuracy and reliability of the encoded adapter function.

[0147] It should be noted that the number of output parameters of the adapter function to be encoded can be one or multiple, and the present invention does not limit this.

[0148] In one example, the data format of the output parameter can be an existing data format. At this time, the data format only includes the data type of the output parameter, such as integer type, floating point type, array, etc. For example, the output parameter can indicate the probability that the image includes the target object, or indicate the feature matrix corresponding to the feature of interest in the image.

[0149] In another example, the data format of the output parameter can also be a custom data format formed by combining and transforming the existing data format. At this time, the data format of the output parameter can include: multiple parameter sub-items that make up the output parameter and the format requirements for each parameter sub-item. For example, the actual output parameter can be an image detection information in a custom data format, and the image detection information includes parameter sub-items such as an image, a detection result, and a detection anchor box. Among them, the format requirement for the image is in the form of an array, and the array includes attributes indicating the storage format of the image and attributes indicating the image size. The format requirement for the detection result is an integer type, specifically 0 or 1. A value of 0 indicates that the image does not include the object to be detected, and a value of 1 indicates that the image includes the object to be detected; the format requirement for the detection anchor box is in the form of an array, and the corresponding array includes the coordinates indicating the left lower end point of the detection anchor box and the coordinates indicating the right upper end point of the detection anchor box.

[0150] In one embodiment, the function input information of the adapter function to be encoded includes: the step output of the previous step associated with the adapter function, and / or, the service data to be processed carried in the service requirement information corresponding to the target service, where the execution order of the previous step is earlier than the execution order of the step corresponding to the adapter function.

[0151] More specifically, when the adapter function to be encoded is not the first step among the multiple steps included in the target service, the function input information of the adapter function to be encoded can be understood as the actual step output of its previous step. For example: when the adapter function to be encoded corresponds to the 3rd step of the target service, the function input information of the adapter function to be encoded can be: the step output obtained after the target service executes the 2nd step, and the step output obtained after the target service executes the 1st step.

[0152] When the adapter function to be encoded is the first step among the multiple steps included in the target service, the function input information of the adapter function to be encoded can be understood as the actual service input of the target service. For example: when the adapter function to be encoded corresponds to the 1st step of the target service, and the target service is used for pollution detection or crack detection of the image to be recognized, the function input information of the adapter function to be encoded can be understood as the input image to be recognized.

[0153] Among them, the previous step associated with the adapter function can be understood as: in the static planning information, the previous step whose step output is used as the function input by the adapter function.

[0154] For example, the service data to be processed carried in the service requirement information corresponding to the target service can be the image to be processed.

[0155] Based on the above settings, compared with the method of specifying the function input constraint method of the adapter function to be encoded in the form of a pattern, by associating the step output of the previous step of the adapter function, and / or, the actual inputs such as the service data to be processed carried in the service requirement information corresponding to the target service, as the function input of the adapter function to be encoded, more concise input prompt information can be provided to reduce the introduction of interference information and improve the accuracy and reliability of the encoded adapter function.

[0156] It should be noted that when specifying the function input of the adapter function to be encoded in the form of a pattern, the function input of the adapter function to be encoded includes at least the parameter name of the corresponding input parameter, the parameter description of the input parameter (which may involve multi-dimensional and multi-type nesting situations), the data type of the input parameter, the value range / value limit of the input parameter, etc. This will make the function input complex and redundant, and inevitably introduce more interference information unrelated to encoding. By using the real input as the function input of the adapter function to be encoded, the function input constraints can be made concise and effective, and the mixing of redundant information can be avoided.

[0157] Of course, it should be noted that it is precisely by leveraging the large-scale heterogeneous computing power resources (such as CPUs, GPUs, and FPGAs) equipped in the intelligent computing center that it is possible to support the rapid processing and analysis of large amounts of data to complete the encoding of the adapter function in an extremely short time. Therefore, it is also possible to support the use of real input as the function constraint of the adapter function to be encoded, that is, while ensuring the accuracy and reliability of the encoded adapter function, the actual execution time of the target service can also be controlled within the expected level, that is, the actual execution time of the target service is less than the set maximum waiting time for the service.

[0158] In one embodiment, the encoding requirement information further includes the function code template of the adapter function to be encoded;

[0159] The function code template of the adapter function to be encoded sequentially includes:

[0160] The header identifier indicating the first digit of the adapter function to be encoded;

[0161] The import code text indicating the function dependencies of the adapter function to be encoded;

[0162] The input parameter code text indicating the function input of the adapter function to be encoded;

[0163] The annotation code text indicating the function annotation of the adapter function to be encoded;

[0164] The function body code text indicating the function functionality of the adapter function to be encoded;

[0165] The output parameter code text indicating the function output of the adapter function to be encoded;

[0166] The tail identifier indicating the last digit of the adapter function to be encoded.

[0167] Based on the above settings, the large model is guided to perform encoding following through a standardized function code template rather than a prompt text, so as to improve the accuracy of the finally obtained adapter function as much as possible.

[0168] Specifically, the function code template of the adapter function to be encoded is used to set the function framework of the adapter function, so as to help the large model build the adapter function by filling in the specific function content within the function framework, thereby increasing the usability of the final adapter function.

[0169] Among them, the head identifier and the tail identifier cooperate with each other to indicate the location of the function code template, helping the function accurately identify the function code template and follow it; the import code text is used to simulate the dependency reference code for encoding the adapter function, so as to guide the large model to import all the required dependency packages at the location indicated by the import code text, to avoid the adapter function reporting an error due to missing dependencies. The input parameter code text is used to simulate the function input parameter code of the adapter function, so as to guide the large model to standardize the input parameters of the adapter function and the data formats of each input parameter. The output parameter code text is used to simulate the function output parameter code of the adapter function, so as to guide the large model to standardize the output parameters of the adapter function and the data formats of each output parameter (in the application, the output parameter code text can be set in the form of a dictionary to help the large model further standardize the function output format following the dictionary). The comment code text is used to simulate the function function comment code of the adapter function, and the function body code text is used to indicate the encoding location of the function body of the adapter function.

[0170] For example, the function code template of the aforementioned adapter function to be encoded can be:

[0171]

[0172]

[0173] In this example, the strings "```python" and "```" can be regarded as the aforementioned head identifier and tail identifier respectively, the string "<import packages>" can be regarded as the import code text, the string "(var_1:dtype_1,var_2:dtype_2)" can be used as the input parameter code text, the string "<function comments:describe function,input and outputs>" can be regarded as the aforementioned comment code text, the string <codes>” can be regarded as the foregoing function body code text, and the string "{{\"output_var_name_1\":output_var_value_1,\"output_var_name_2\":output_var_value_2}}" can be regarded as the foregoing output parameter code text.

[0174] In one embodiment, the encoding requirement information further includes: a plurality of input parameter names and a plurality of output parameter names;

[0175] Wherein, the plurality of input parameter names are used to indicate: a plurality of input parameters corresponding to the adapter function in the static planning information;

[0176] The plurality of output parameter names are used to indicate: a plurality of output parameters corresponding to the adapter function in the static planning information.

[0177] Based on the above settings, by concisely indicating all the input parameters and all the output parameters of the adapter function to be encoded, it is possible to avoid the omission of output parameters in the function output mode of the adapter function to be encoded, and to avoid the omission of input parameters in the function input information of the adapter function to be encoded, ensuring that the input parameters and output parameters in the adapter function conform to the indication of the corresponding step data body in the static planning information.

[0178] In one example, all the input parameters and all the output parameters of the adapter function to be encoded can be indicated by tag [INPUT_NAMES] and [OUTPUT_NAMES], wherein INPUT_NAMES is used to display the foregoing plurality of input parameter names, and OUTPUT_NAMES is used to display the foregoing plurality of output parameter names.

[0179] In one embodiment, the step S2 includes:

[0180] Step S21: Using the encoding requirement information as an encoding prompt (prompt), based on the computing power of the intelligent computing center, iteratively execute the function encoding task;

[0181] The nth iteration in the iteration includes:

[0182] Step S211: When the function encoding task in the nth iteration is successfully executed, form the adapter function based on the execution result of the function encoding task in the nth iteration, and exit the iteration;

[0183] Step S212: When the function encoding task in the nth iteration fails and n is less than or equal to the set iteration count threshold, use the task failure information of the function encoding task in the nth iteration and the encoding requirement information as encoding prompt words, and use the computing power of the intelligent computing center to execute the function encoding task in the (n + 1)th iteration;

[0184] Step S213: When the function encoding task in the nth iteration fails and n is greater than the iteration count threshold, exit the iteration and output the task failure information of the function encoding task in the nth iteration;

[0185] Wherein, n is an integer greater than or equal to 1, and when n = 1, the function encoding task in the nth iteration uses the encoding requirement information as the encoding prompt word.

[0186] Based on the above settings, it supports the large model to iteratively execute the function encoding task within the set iteration count, thereby increasing the probability of obtaining a reliable adapter function.

[0187] Exemplarily, it can be set that when the function encoding task in the nth iteration is executed without error, it is determined that the function encoding task in the nth iteration is successfully executed; otherwise, it is determined that the function encoding task in the nth iteration fails.

[0188] The task failure information of the function encoding task in the nth iteration at least includes the execution error information of the function encoding task in the nth iteration. Among them, the foregoing execution error information can be further segmented to filter out redundant error data (referring to error nesting content unrelated to the execution of the function encoding task) in the execution error information, so as to streamline the data in the task failure information, help the large model understand the encoding error, and avoid the mixing of unnecessary interference information, thereby increasing the probability of successful encoding by the large model during iteration; in addition, the error code text corresponding to the execution error information can also be supplemented in the task failure information to help the large model obtain more background knowledge of the current error situation, further increasing the probability of successful encoding by the large model during iteration.

[0189] Please refer to Figure 4 , Figure 4 which is a schematic structural diagram of a device for generating a service execution component through the computing power of an intelligent computing center provided by the present invention. As Figure 4 shown, the device 400 for generating a service execution component through the computing power of an intelligent computing center includes:

[0190] A requirements acquisition module 401, configured to obtain encoded requirements information from static planning information based on the computing power of an intelligent computing center, where the static planning information is used to indicate the service processing logic of a target service, and the encoded requirements information includes: the function name of an adapter function to be encoded, the step description text of the step corresponding to the adapter function to be encoded in the static planning information, the function output path of the adapter function to be encoded, the function output mode of the adapter function to be encoded, and the function input information of the adapter function to be encoded. The target service includes a vision processing service based on a computer vision agent, and the encoded requirements information is used to represent the personalized function requirements in the vision processing service;

[0191] An encoding module 402, configured to use the encoded requirements information as an encoding prompt word and encode an adapter function based on the computing power of an intelligent computing center, where the adapter function is used to cooperate with a preset basic function to execute the target service, the basic function is used to implement the general functions in the vision processing service, and the adapter function is used to implement the personalized functions in the vision processing service.

[0192] In one embodiment, the function output mode of the adapter function to be encoded does not include the parameter explanation text of the output parameter of the adapter function to be encoded;

[0193] The function output mode of the adapter function to be encoded includes at least one of the following:

[0194] The parameter name of the output parameter of the adapter function to be encoded;

[0195] The data format of the output parameter of the adapter function to be encoded;

[0196] The value range of the output parameter of the adapter function to be encoded.

[0197] In one embodiment, the function input information of the adapter function to be encoded includes: an input parameter value dictionary of the step corresponding to the adapter function to be encoded in the static planning information, and the input parameter value dictionary is used to indicate the storage location of the image and / or text to be processed during the execution stage of the vision processing service.

[0198] In one embodiment, the encoded requirements information further includes a function code template of the adapter function to be encoded;

[0199] The function code template of the adapter function to be encoded sequentially includes:

[0200] A head identifier indicating the first digit of the adapter function to be encoded;

[0201] A package import code text indicating the function dependencies of the adapter function to be encoded;

[0202] The input parameter code text of the function input that indicates the adapter function to be encoded;

[0203] The annotation code text of the function annotation that indicates the adapter function to be encoded;

[0204] The function body code text of the function function that indicates the adapter function to be encoded;

[0205] The output parameter code text of the function output that indicates the adapter function to be encoded;

[0206] The tail identifier at the end that indicates the adapter function to be encoded.

[0207] In one embodiment, the encoding requirement information further includes: a plurality of input parameter names and a plurality of output parameter names;

[0208] Wherein, the plurality of input parameter names are used to indicate: the plurality of input parameters corresponding to the adapter function in the static planning information;

[0209] The plurality of output parameter names are used to indicate: the plurality of output parameters corresponding to the adapter function in the static planning information.

[0210] In one embodiment, the encoding module 402 is specifically configured to:

[0211] Using the encoding requirement information as an encoding prompt word, based on the computing power of the intelligent computing center, iteratively execute the function encoding task;

[0212] The nth iteration in the iteration includes:

[0213] When the function encoding task in the nth iteration is successfully executed, form the adapter function based on the execution result of the function encoding task in the nth iteration, and exit the iteration;

[0214] When the function encoding task in the nth iteration fails and n is less than or equal to the set iteration times threshold, using the task failure information of the function encoding task in the nth iteration and the encoding requirement information as an encoding prompt word, use the computing power of the intelligent computing center to execute the function encoding task of the n+1th iteration;

[0215] When the function encoding task in the nth iteration fails and n is greater than the iteration times threshold, exit the iteration, and output the task failure information of the function encoding task in the nth iteration;

[0216] Wherein, n is an integer greater than or equal to 1, and when n = 1, the function encoding task in the nth iteration uses the encoding requirement information as an encoding prompt word.

[0217] Schematic rather than restrictive, those of ordinary skill in the art, inspired by the present invention and without departing from the spirit of the present invention and the scope protected by the claims, can also make many forms, all of which fall within the protection of the present invention.

[0218] The device for generating service execution components through the computing power of the intelligent computing center provided by the present invention can implement each process of the above-mentioned method for generating service execution components through the computing power of the intelligent computing center. The technical features correspond one by one and can achieve the same technical effects. To avoid repetition, they will not be elaborated here.

[0219] It should be noted that the device for generating service execution components through the computing power of the intelligent computing center in the present invention can be a device, or a component, integrated circuit, or chip in an electronic device.

[0220] The present invention also provides an electronic device. Refer to Figure 5 , Figure 5 is a schematic structural diagram of an electronic device provided in an embodiment of the present invention. The electronic device includes a memory 501, a processor 502, and a program or instruction running on the memory 501. When the program or instruction is executed by the processor 502, it can implement Figure 1 any step in the corresponding method embodiment for generating service execution components through the computing power of the intelligent computing center and achieve the same beneficial effects. They will not be elaborated here.

[0221] Among them, the processor 502 can be a CPU, ASIC, FPGA, or GPU.

[0222] Those of ordinary skill in the art can understand that all or part of the steps of implementing the above-mentioned method embodiment for generating service execution components through the computing power of the intelligent computing center can be completed by hardware related to program instructions, and the program can be stored in a readable medium.

[0223] The present invention also provides a readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it can implement the above-mentioned Figure 3 any step in the corresponding method embodiment for generating service execution components through the computing power of the intelligent computing center and can achieve the same technical effects. To avoid repetition, they will not be elaborated here. The storage medium such as a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disc, etc.

[0224] The terms "first", "second", etc. in the present invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. In addition, the terms "comprising", "having" and any of their variants are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device 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 not clearly listed or inherent to these processes, methods, products or devices. In addition, the use of "and / or" in the present invention means at least one of the connected objects. For example, A and / or B and / or C means including the 7 cases of A alone, B alone, C alone, A and B both present, B and C both present, A and C both present, and A, B and C all present.

[0225] It should be noted that in this text, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, article or device comprising that element.

[0226] Through the description of the above embodiments, those skilled in the art can clearly understand that the above-described embodiment methods can be implemented by means of software plus a necessary general hardware platform. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disc) and includes several instructions for causing a terminal (which can be a mobile phone, computer, server, air conditioner, or a second terminal device, etc.) to execute the methods of the various embodiments of the present invention.

[0227] The above describes the embodiments of the present invention in conjunction with the drawings, but the present invention is not limited to the above specific embodiments. The above specific embodiments are merely illustrative and not restrictive. Under the inspiration of the present invention, those of ordinary skill in the art can also make many forms without departing from the purpose of the present invention and the scope protected by the claims, and all of them belong to the protection scope of the present invention.< / codes>

Claims

1. A method for generating business execution components through the computing power of an intelligent computing center, characterized in that Including: Step S1: Obtain coding requirement information from static planning information based on the computing power of the intelligent computing center. The static planning information is used to indicate the business processing logic of the target business. The coding requirement information includes: the function name of the adapter function to be coded, the step description text of the step corresponding to the adapter function to be coded in the static planning information, the function output path of the adapter function to be coded, the function output mode of the adapter function to be coded, and the function input information of the adapter function to be coded. The target business includes a vision processing business based on a computer vision agent, and the coding requirement information is used to represent the personalized function requirements in the vision processing business. Step S2: Use the coding requirement information as coding prompt words and code an adapter function based on the computing power of the intelligent computing center. The adapter function is used to cooperate with a preset basic function to execute the target business. The basic function is used to implement the general functions in the vision processing business, and the adapter function is used to implement the personalized functions in the vision processing business.

2. The method according to claim 1, wherein The function output mode of the adapter function to be coded does not include the parameter explanation text of the output parameters of the adapter function to be coded. The function output mode of the adapter function to be coded includes at least one of the following: The parameter name of the output parameter of the adapter function to be coded; The data format of the output parameter of the adapter function to be coded; The value range of the output parameter of the adapter function to be coded.

3. The method according to claim 1, wherein The function input information of the adapter function to be coded includes: the step output of the previous step associated with the adapter function, and / or, the business data to be processed carried in the business requirement information corresponding to the target business, where the execution order of the previous step is earlier than the execution order of the step corresponding to the adapter function.

4. The method according to claim 1, wherein The coding requirement information further includes the function code template of the adapter function to be coded. The function code template of the adapter function to be coded sequentially includes: The head identifier indicating the first digit of the adapter function to be coded; The import package code text indicating the function dependencies of the adapter function to be coded; The input parameter code text indicating the function input of the adapter function to be coded; The annotation code text indicating the function annotation of the adapter function to be coded; The function body code text indicating the function function of the adapter function to be coded; The output parameter code text indicating the function output of the adapter function to be coded; The tail identifier indicating the last digit of the adapter function to be coded.

5. The method according to claim 1, wherein The coding requirement information further includes: multiple input parameter names and multiple output parameter names; Among them, the multiple input parameter names are used to indicate: multiple input parameters corresponding to the adapter function in the static planning information; The multiple output parameter names are used to indicate: multiple output parameters corresponding to the adapter function in the static planning information.

6. The method according to any one of claims 1-5, characterized in that, Step S2 includes: Step S21: Use the coding requirement information as coding prompt words and iteratively execute the function coding task based on the computing power of the intelligent computing center. The nth iteration in the iteration includes: Step S211: When the function encoding task in the nth iteration is successfully executed, form the adapter function based on the execution result of the function encoding task in the nth iteration, and exit the iteration; Step S212: When the function encoding task in the nth iteration fails and n is less than or equal to the set iteration times threshold, use the task failure information of the function encoding task in the nth iteration and the encoding requirement information as encoding prompt words, and use the computing power of the intelligent computing center to execute the function encoding task in the (n + 1)th iteration; Step S213: When the function encoding task in the nth iteration fails and n is greater than the iteration times threshold, exit the iteration and output the task failure information of the function encoding task in the nth iteration; Wherein, n is an integer greater than or equal to 1, and when n = 1, the function encoding task in the nth iteration uses the encoding requirement information as the encoding prompt word.

7. An apparatus for generating a service execution component by the computing power of an intelligent computing center, characterized in that, It includes: A requirement acquisition module, configured to obtain encoding requirement information from static planning information based on the computing power of the intelligent computing center, wherein the static planning information is used to indicate the business processing logic of the target business, and the encoding requirement information includes: the function name of the adapter function to be encoded, the step description text of the corresponding step of the adapter function to be encoded in the static planning information, the function output path of the adapter function to be encoded, the function output mode of the adapter function to be encoded, the function input information of the adapter function to be encoded, the target business includes a visual processing business based on a computer vision agent, and the encoding requirement information is used to represent the personalized function requirements in the visual processing business; An encoding module, configured to use the encoding requirement information as an encoding prompt word and encode to obtain an adapter function based on the computing power of the intelligent computing center, wherein the adapter function is used to cooperate with a preset basic function to execute the target business, the basic function is used to implement the general functions in the visual processing business, and the adapter function is used to implement the personalized functions in the visual processing business.

8. A server, characterized in that, It includes: A processor, a memory, and a program stored on the memory and executable on the processor. When the program is executed by the processor, it implements the steps of the method for generating a service execution component through the computing power of the intelligent computing center as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, A computer program is stored on the computer-readable storage medium. When the computer program is executed by the processor, it implements the steps of the method for generating a service execution component through the computing power of the intelligent computing center as described in any one of claims 1 to 6.

10. A computer program product, characterized in that, It includes computer instructions. When the computer instructions are executed by the processor, it implements the steps of the method for generating a service execution component through the computing power of the intelligent computing center as described in any one of claims 1 to 6.

Images