Computing power information generation method and device based on intelligent computing center

Abstract

The invention provides a computing power information generation method and device based on 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, obtaining a first text, a second text and a third text of a target function, the target function is a basic function for realizing a general function in the target service, the first text is used for representing a function code of the target function, the second text is used for representing a function of the target function, and the third text is used for representing application information of the target function; and S2, the first text, the second text and the third text are fused, function information of the target function is obtained, and the function information is used for being loaded by a function manager so as to import and run a function code of the target function. Through centralized management of the function code, the function and the application information of the target function, the information management effect of a plurality of basic functions can be remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical fields of intelligent computing centers, intelligent computing centers, and computing power infrastructure technologies, and particularly relates to a method and device for generating information based on 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, to mainly provide the required computing power, data, and algorithms for artificial intelligence applications (such as scenarios like artificial intelligence deep learning model development, model training, and model inference). An intelligent computing center covers facilities, hardware, and software, and can provide full-stack capabilities from underlying computing power to top-level application enabling.

[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, and mainly provides services to society through computing power infrastructure.

[0007] Currently, how to quickly generate an intelligent agent (agent) that matches the requirements is an urgent problem to be solved. A possible solution is: by means of custom coding, a batch of general functions (function) and the function patterns corresponding to the functions are pre-constructed. This part of the pre-constructed functions can be called base functions (base function), and the function patterns corresponding to the functions can be called base function schemas (base function schema). By providing the base function schema to the large model so that the large model masters the uses and usages of the base function, the large model, as the business processing core of the pre-generated intelligent agent, can, after the intelligent agent receives the business requirements input by the user, dynamically form a static plan (plan) that meets the above business requirements based on the mastered uses and usages of the base function, and execute the static plan to achieve the response to the business requirements.

[0008] In the application, it is found that the code of the basic function and the information related to the basic function (such as the basic function mode, etc.) are scattered in different storage locations, which makes the information management of the basic function very chaotic.

[0009] It can be seen that based on the information management method in the prior art, there is a problem of very poor information management effect in managing the code of the basic function and its related information. Summary of the Invention

[0010] The purpose of the present invention is to provide an information generation method and device based on the computing power of an intelligent computing center, which are used to solve the problem of very poor information management effect existing in managing the code of the basic function and its related information in the prior art.

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

[0012] In a first aspect, the present invention provides an information generation method based on the computing power of an intelligent computing center, and the method includes:

[0013] Step S1, obtain a first text, a second text, and a third text of a target function, where the target function is one of a plurality of basic functions, and the plurality of basic functions are used to implement general functions in a target service, the target service includes a visual processing service based on a computer vision agent, the first text is used to represent the function code of the target function, the second text is used to represent the function function of the target function, the third text is used to represent the application information of the target function, and the application information of the target function includes: the usage log of the target function and the function encapsulation information of the target function, and the usage log of the target function includes service data processed by the target function;

[0014] Step S2, fuse the first text, the second text, and the third text to obtain function information of the target function, where the function information is used to be loaded by a function manager to import and run the function code of the target function.

[0015] In an embodiment, the second text includes at least one of the following:

[0016] A first text segment, and the version type of the first text segment is a natural language version;

[0017] A second text segment, and the version type of the second text segment is a function mode version in an unmasked state;

[0018] A third text segment, and the version type of the third text segment is a function mode version in a masked state;

[0019] Among them, the first text segment, the second text segment, and the third text segment are all used to represent the function function of the target function.

[0020] In one embodiment, when the second text includes the third text segment, the function manager stores mask mapping information, and the mask mapping information is used to restore the third text segment to the second text segment.

[0021] In one embodiment, the target function is one function version among all function versions corresponding to a first basic function, and the first basic function is one basic function among the multiple basic functions.

[0022] In one embodiment, step S2 includes:

[0023] Step S21: Concatenate the first text, the second text, and the third text to obtain the function information of the target function.

[0024] In one embodiment, after step S2, the method further includes:

[0025] Step S3: Receive an information update instruction, where the information update instruction is used to update the function information of the target function;

[0026] Step S4: In response to the information update instruction, update the function information of the target function to obtain the updated information of the target function;

[0027] Among them, the second text in the updated information is different from the second text in the function information, and / or the third text in the updated information is different from the third text in the function information.

[0028] In a second aspect, the present invention also provides an information generation device based on the computing power of an intelligent computing center, including:

[0029] A text acquisition module, configured to acquire a first text, a second text, and a third text of a target function, where the target function is one of multiple basic functions, the multiple basic functions are used to implement general functions in a target service, the target service includes a vision processing service based on a computer vision agent, the first text is used to represent the function code of the target function, the second text is used to represent the function function of the target function, the third text is used to represent the application information of the target function, and the application information of the target function includes: the usage log of the target function and the function encapsulation information of the target function, and the usage log of the target function includes service data processed by the target function;

[0030] An information generation module, configured to fuse the first text, the second text, and the third text to obtain function information of the target function, where the function information is used to be loaded by a function manager to import and run function code of the target function.

[0031] 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, where when the program is executed by the processor, the steps in the information generation method based on the computing power of an intelligent computing center as described in the first aspect above are implemented.

[0032] 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, the steps in the information generation method based on the computing power of an intelligent computing center as described in the first aspect above are implemented.

[0033] 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, the steps in the information generation method based on the computing power of an intelligent computing center as described in the first aspect above are implemented.

[0034] In the present invention, by acquiring and fusing the first text, the second text, and the third text of the target function, centralized management of the function code, function function, and application information of the target function is achieved, avoiding the management chaos problem caused by decentralized storage of the code of basic functions and their related information, and significantly improving the information management effect of multiple basic functions. Description of the Drawings

[0035] By reading the detailed description of the preferred embodiments below, 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:

[0036] Figure 1 is an architecture diagram of a two-stage agent automation workflow provided by the present invention;

[0037] Figure 2 is a schematic diagram of the results of an experimental agent automation workflow using Qwen2.5 - 72B - Instruct provided by the present invention;

[0038] Figure 3 is a flowchart of an information generation method based on the computing power of an intelligent computing center provided by the present invention;

[0039] Figure 4It is a schematic structural diagram of an information generation device that utilizes the computing power of an intelligent computing center provided by the present invention;

[0040] Figure 5 It is a schematic structural diagram of an electronic device provided by the present invention. Detailed implementation manners

[0041] Next, the technical solutions in the present invention will be clearly and completely described in conjunction with the accompanying drawings in the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts fall within the protection scope of the present invention.

[0042] First, the technical terms related to the present invention will be briefly explained below.

[0043] The "computing power" referred to in the present invention is the ability of a computer device or a computing / data center to process information, which 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 computing power infrastructure.

[0044] The "computational power" (Computational Power, CP) referred to in the present invention is a kind of ability of the data center server to process data and achieve the 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 .

[0045] The "carrying capacity" (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 including network architecture, network bandwidth, transmission delay, intelligent management and scheduling, etc. The carrying capacity 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 carrying capacity adopts the video memory bandwidth.

[0046] The "Storage Power" (SP) described in the present invention is the comprehensive ability of a data center in four aspects: data storage capacity, performance, security and reliability, and green and low-carbon. It is a comprehensive indicator to measure the data storage ability of a data center, including external storage devices such as storage arrays and built-in storage devices of servers. 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 / write operations per second per unit capacity (Input / Output Operations Per Second / TB, IOPS / TB), and the disaster recovery ratio is an important manifestation of security and reliability.

[0047] 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 power, and can realize the centralized computing, storage, transmission, and application of information.

[0048] 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, and satellite Internet, computing power infrastructures such as data centers, general computing power centers, intelligent computing centers, and supercomputing centers, and new technology facilities such as artificial intelligence, blockchain, and quantum computing.

[0049] The "computing power" described in the present invention includes general computing power, intelligent computing power, and super computing power.

[0050] The "general computing power" described in the present invention is the computing ability 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.

[0051] The "intelligent computing power" described in the present invention is for various artificial intelligence innovation applications, and is a computing platform based on the large-scale deployment of dedicated chips such as graphics processing unit (GPU), field programmable gate array (FPGA), and application specific integrated circuit (ASIC), such as natural language processing and machine vision.

[0052] The "super computing power" described in the present invention is mainly the computing ability 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, and is mainly used for computing in cutting-edge scientific fields, such as planetary simulation, drug molecule design, and gene analysis.

[0053] The "Intelligent Computing Center" described in the present invention refers to a facility that uses large-scale heterogeneous computing power resources, including general computing power (CPU) and intelligent computing power (GPU, FPGA, etc.), and mainly provides the required computing power, data, and algorithms for artificial intelligence applications (such as scenarios like artificial intelligence deep learning model development, model training, and model inference). The intelligent computing center covers facilities, hardware, and software, and can provide full-stack capabilities from underlying computing power to top-level application enabling.

[0054] The "Intelligent Computing Center" described in the present invention includes, but is not limited to, the "Intelligent Computing Center".

[0055] The "Intelligent Computing Center" described in the present invention, that is, the artificial intelligence computing center, is a type of computing power 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.

[0056] 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, and having computing power, carrying capacity, and storage capacity, including general data centers, intelligent computing centers, supercomputing centers, etc.

[0057] The "Supercomputing Center" described in the present invention, that is, the supercomputing data center, 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.

[0058] The "Computing Power Resources" described in the present invention refer to technologies and facilities required for the development of the digital society and having information computing, transmission, storage, and application capabilities, including, but not limited to, computing resources such as CPU and GPU, 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.

[0059] The "models" and "large models" described in the present invention include, but are not limited to, "large language models" and "multimodal large models".

[0060] The "Large Language Model" described in the present invention refers to a large language model (LLM), which is a language model with a relatively large number of parameters, aiming to understand and generate human language, trained through a large amount of text data, and can perform a wide range of tasks including text summarization, translation, sentiment analysis, etc.

[0061] The "multimodal large models" referred to in the present invention means: models trained by jointly using multimodal information such as text, images, videos, and audio, including but not limited to multimodal large language models.

[0062] The "function information" referred to in the present invention means: the set of all data associated with a basic function, including but not limited to: the function code of the basic function, the basic function mode of the basic function, the function encapsulation information of the basic function, etc. In applications, the above "function information" usually exists in the form of an object, and a type of object used to refer to "function information" can be defined as a FuncInfoBuilder object.

[0063] The "application information" referred to in the present invention means: all other data in all the data associated with the basic function except the function code and the function function of the basic function.

[0064] The "usage log" referred to in the present invention means: the operation log of a basic function, including but not limited to: the time when the basic function is called, the business requirements corresponding to the period when the basic function is called, the data to be processed during the period when the basic function is called, the function result output during the period when the basic function is called, the error message generated when the basic function is not called normally, etc.

[0065] The "function encapsulation information" referred to in the present invention means: the encapsulation transformation supported by a basic function, including but not limited to: the transformation of the function input of the basic function (such as adjusting the number of original function input parameters of the basic function, parameter types, parameter value constraints, etc.), the transformation of the function output of the basic function (such as adjusting the number of original function input parameters of the basic function, parameter types, parameter value constraints, etc.).

[0066] The "information management effect" referred to in the present invention means: the usability of all data associated with a basic function, including but not limited to: the read and write rate of data, the query efficiency of data, the readability of data.

[0067] The "agent" referred to in the present invention is an agent 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 learned by itself, and then executes actions to affect the environment or achieve a predetermined goal. Agents are widely used in the field of artificial intelligence, and are commonly found in automated systems, robots, virtual assistants, and game characters, etc. The core lies in the ability to learn independently and evolve continuously to better complete tasks and adapt to complex environments.

[0068] The "base function" described in the present invention refers to a function pre-constructed by a developer according to the general / high-frequency business operations of the business scenario faced by the agent. For example, when the agent is specifically a computer vision agent, the general / high-frequency business operations of the business scenario faced by the agent may include: image preprocessing operations, image segmentation operations, image noise filtering operations, image target detection operations, etc.

[0069] The "base function schema" described in the present invention refers to the "base function usage manual" formed by the developer based on the functions and usage of the base function. The base function schema may include items such as name, description, type, etc. Among them, the name item may 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 function through the "base function schema".

[0070] 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 function 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. 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).

[0071] The "autocoder" described in the present invention refers to a large model for code writing, which is used to write adapter functions.

[0072] 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.

[0073] The "step data body" described in the present invention is a data representation corresponding to a business processing step, including but not limited to: the step description text of a 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 understanding and implementation by users or developers), the step number of a corresponding business processing step among the multiple business processing steps, the input data of a 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 a corresponding business processing step (specific methods are called through method names).

[0074] The "pipeline" described in the present invention refers to the process of executing a static plan.

[0075] The "text" described in the present invention refers to a set formed by several data, for example: a character set formed by code characters.

[0076] The "splicing" described in the present invention refers to the operation of placing the text to be spliced completely within the information formed after splicing, that is, the data content of the text before and after splicing does not change at all, to ensure the integrity of the semantics of the text after splicing. For multiple texts to be spliced, they can be spliced in a head-to-tail connection manner, or placed at the set positions in a blank template according to set rules to complete the splicing, or completed by referring to the object instantiation method (the information after splicing is used as the instantiated object, and the text to be spliced is used as the attribute of the instantiated object).

[0077] Regarding the problem of quickly generating an agent that matches user needs, the present invention provides an agent automation workflow solution, which is specifically as follows: By means of custom coding, a batch of general functions and the corresponding function patterns of the functions are pre-constructed. 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 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 the actual scenario, the base functions may not cover all business needs. For this problem, the large model can write code temporarily to generate adapter functions that can fill the gaps in the base functions to cooperate with the base functions to achieve full coverage of the business needs.

[0078] In one example, the above-mentioned agent automation workflow solution is implemented by using the powerful computing power and flexible resource configuration of the intelligent computing center. The corresponding architecture of the two-stage auto-pipeline can be as Figure 1 shown Figure 1 The relevant illustrations in

[0079] One user query: Users input their requests or demand content.

[0080] 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.

[0081] Plan list: A list of specific steps generated according to user needs.

[0082] Retrieve all-step function dependencies: Retrieve the function dependencies related to each step to ensure that all required functions are available.

[0083] Get executable input values:

[0084] Given values: Values directly provided to a function.

[0085] User inputs: Dynamic input values provided by the user in a query.

[0086] Previous function outputs: Function output values from previous steps.

[0087] Media download: This step can also support downloading additional required media content.

[0088] Execute one step: Refers to the indication of the execution details of a specific step.

[0089] Execute function: Used to actually call a function to perform data processing work based on the obtained input values.

[0090] Base function: Performs specific basic functions to support the agent's set basic business processing capabilities.

[0091] Adapter function (autoencoder): Serves as a supplement to the base function.

[0092] Retrieve depended function schemas: Obtain the structures or schemas of other dependent functions.

[0093] Autoencoder coding: Automatically generate the code involved in the above process.

[0094] Exception code: If an error or exception occurs, the system will return a specific exception code for subsequent analysis and debugging. For example:

[0095] Exceptions can be classified into the following 6 categories:

[0096] 0: No exception, pipeline execution successful

[0097] 1: Other pipeline exceptions

[0098] 2: Autoencoder execution failed

[0099] 3: Base function execution failed

[0100] 4: Plan retrieval failed (usually due to failure to convert plan to json)

[0101] 5: The placeholder expression of the variable in the plan is incorrect and the parsing fails.

[0102] It should be noted that the application of the two-stage auto-pipeline has at least the following advantages:

[0103] 1. Higher accuracy: There are more cases where the given plan can be successfully executed;

[0104] 2. By decomposing tasks, the total inference time of the LLM can be shortened;

[0105] 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.

[0106] The results of experimenting with the two-stage auto-pipeline using the Qwen2.5-72B-Instruct tool for different data sources are as Figure 2 shown.

[0107] 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, confusing use of picture lists and video paths, 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 handling of complex parameters, etc. In addition, some errors were caused by poor compliance with the json format of the plan output.

[0108] Based on the information management method in the prior art, managing the code of the basic function and its related information, the code of the basic function and the information related to the basic function (such as the basic function mode, etc.) will be scattered in different storage locations, which makes the information management effect of the basic function very poor.

[0109] Based on this, the present invention provides an information generation method based on the computing power of an intelligent computing center. By obtaining and fusing the first text, the second text, and the third text of the target function, centralized management of the function code, function function, and application information of the target function is achieved, avoiding the management chaos problem caused by scattered storage of the code of the basic function and its related information, and significantly improving the information management effect of multiple basic functions.

[0110] Please refer to Figure 3 , Figure 3 which is an information generation method based on the computing power of an intelligent computing center provided by the present invention. As Figure 3 shown, it includes the following steps:

[0111] Step S1: Obtain the first text, second text, and third text of the target function.

[0112] Among them, the target function is one of multiple basic functions, and the multiple basic functions are used to implement general functions in the target business. The target business includes vision processing services based on computer vision agents. The first text is used to represent the function code of the target function, the second text is used to represent the function function of the target function, and the third text is used to represent the application information of the target function. The application information of the target function includes: the usage log of the target function and the function encapsulation information of the target function. The usage log of the target function includes the business data processed by the target function.

[0113] Exemplarily, the vision processing service based on computer vision agents can 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.

[0114] It should be noted that the first text is the text representation of the function code of the target function. That is, the process of reading the first text can be regarded as the process of importing the function code of the target function.

[0115] The second text is provided to the task planning large model to teach the task planning large model to master the use and usage of the target function. It should be understood that the second text can be a generalization description of the function function of the target function (such as the target function is an image denoising function, the required image format of the function input is XX, the image format of the function output is YY, and the image denoising algorithm used in the image denoising process is ZZ), or a specific description of the function function of the target function corresponding to the planning data for fine-tuning the task planning large model (such as the target function is used to perform image denoising processing, the function input is image A, the function output is image B, and a convolutional neural network is used for denoising processing during the image denoising process).

[0116] The third text is the text representation of other data in all the data associated with the target function except the first text and the second text.

[0117] Exemplarily, the usage log of the target function can include: the time when the target function is called, the business requirements corresponding to the period when the target function is called, the data to be processed during the period when the target function is called (that is, the business data processed by the target function, such as data like images, image blocks, image features, etc.), the function results output during the period when the target function is called, the error information generated when the target function is not called normally, etc.

[0118] Exemplarily, the function encapsulation information of the target function may include: transformation of the function input of the target function (such as adjusting the number of function input parameters of the original target function, parameter types, parameter value constraints, etc.), and transformation of the function output of the target function (such as adjusting the number of function input parameters of the original target function, parameter types, parameter value constraints, etc.).

[0119] Through the above encapsulation transformation process, by using the flexibly adjustable function input and / or function output, the reliability of data flow between the target function and other functions (such as basic functions / adapter functions) is improved, and further the execution success probability of the task planning based on the target function is enhanced.

[0120] Step S2: Fuse the first text, the second text, and the third text to obtain the function information of the target function.

[0121] Among them, the function information is used to be loaded by the function manager to import and run the function code of the target function.

[0122] Exemplarily, the function manager can be called FuncDataManager.

[0123] The function manager encapsulates the function running method of the target function. When the function manager reads the first text in the function information, it can realize the import of the target function, and under the condition of meeting the set conditions, call the aforementioned function running method to start the target function.

[0124] Among them, the set conditions may be: receiving a function running instruction corresponding to the target function sent by the planning executor; the planning executor is used to execute the task planning output by the task planning large model.

[0125] It should be noted that to avoid the usage log of the target function occupying too much storage space, it can be set to periodically clean the usage log of the target function to delete the log data in the usage log that is too far from the current time (such as more than one week).

[0126] The present invention realizes the centralized management of the function code, function function, and application information of the target function by obtaining and fusing the first text, the second text, and the third text of the target function, avoids the management chaos problem caused by the scattered storage of the code and its related information of the basic functions, and significantly improves the information management effect of multiple basic functions.

[0127] It should be noted that the centralized management of the function code, function function, and application information of the target function can significantly improve the readability of data. Specifically, for users who are not familiar with basic functions, they can learn the function code, function function, and application information of the target function only through the function information, without the need for users to search for other files additionally, and there will be no data omission problem, which can greatly increase the efficiency and effect of users' mastery / learning of the target function.

[0128] For users who are familiar with basic functions, based on the above settings, for all management and call operations of the target function, such users can only carry out operations facing the function information, which greatly facilitates the operation of the target function by users. Moreover, since all data associated with the target function is recorded in the function information, the risk of inputting duplicate data / error data can be reduced, which greatly improves the operation accuracy of all data associated with the target function.

[0129] Based on the above settings, on the premise of ensuring the accuracy of information management, it is possible to support flexible information management operations for each of multiple basic functions, such as configuring one or more function policy data unique to each basic function (such as masks for basic function modes, prompt words for basic function models, constraint descriptions for basic function models, etc.) to adapt to the application requirements of different function functions corresponding to each basic function, helping the task planning large model better master the functions and usage methods of each basic function, and helping the planning executor better complete the scheduling of each basic function.

[0130] Moreover, since the function code is also stored in the function information, when the function manager reads the first text in the function information, the import of the target function can be realized, which can improve the import efficiency of the target function to a certain extent.

[0131] In one embodiment, the second text includes at least one of the following:

[0132] A first text segment, and the version type of the first text segment is the natural language version;

[0133] A second text segment, and the version type of the second text segment is the function mode version in the unmasked state;

[0134] A third text segment, and the version type of the third text segment is the function mode version in the masked state;

[0135] Among them, the first text segment, the second text segment, and the third text segment are all used to represent the function function of the target function.

[0136] The natural language version can be understood as a text format in pure text form. For example, when the target function is an image denoising function, the first text segment can be "The main function of the image denoising function is to remove or reduce noise in the image, thereby improving the quality and visualization effect of the image. The function can identify different types of noise, such as Gaussian noise, salt-and-pepper noise, and Poisson noise, in order to select an appropriate denoising algorithm for processing. It implements a variety of denoising algorithms, including mean filtering, median filtering, Gaussian filtering, bilateral filtering, and non-local means, etc. Users can flexibly select according to specific needs. The function also allows users to adjust algorithm parameters in order to achieve the best balance between denoising effect and image detail retention. Especially in Gaussian filtering, the standard deviation of the filter can be adjusted to optimize the result. In addition, the image denoising function aims to remove noise while trying to retain the details and edges of the image, avoiding over-smoothing that causes image blurring. To improve the user experience, the function supports input and output of multiple image formats and provides a comparison of the images before and after denoising to help users visually evaluate the effect. To improve processing efficiency, the function also supports batch denoising of multiple images and provides a user-friendly operation interface for users to set. Finally, the function can also output image quality evaluation metrics, such as peak signal-to-noise ratio and structural similarity index, to help users quantify the denoising effect, so as to better meet practical applications".

[0137] The function mode version can be understood as a text format that meets the requirements of the json schema, including multiple identification areas, and corresponding data is set in each identification area, such as items like 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 feature 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.

[0138] Among them, the function mode version in the masked state can be understood as: a text format that has been masked and meets the requirements of the json schema; similarly, the function mode version in the unmasked state can be understood as: a text format that has been masked and meets the requirements of the json schema.

[0139] Among them, the masking process can be understood as: an operation of replacing the parameter names and / or function names included in the name item with target characters, where the target characters are used to suppress or even avoid the over-association of the function feature by the task planning large model based on the function name, so as to improve the fine-tuning effect of the task planning large model during the fine-tuning process.

[0140] In this embodiment, by providing multiple versions of text segments for representing the function capabilities of the target function, more flexible forms of function capabilities are supported, thereby meeting the application requirements in different scenarios and making the application of the present invention more flexible and convenient.

[0141] In one embodiment, when the second text includes the third text segment, the function manager stores mask mapping information, which is used to restore the third text segment to the second text segment.

[0142] Among them, the mask mapping information records the data mapping relationship between the target character and the parameter name and / or function name replaced by the target character during the aforementioned mask processing process.

[0143] In this embodiment, the function manager performs unified mask management on multiple basic functions to avoid conflicts in the target characters used by different basic functions, which can further improve the information management effect of multiple basic functions.

[0144] It should be noted that when the second text includes the third text segment, for actual business requirements, the planning data output by the task planning large model will include mask content. At this time, the function manager performs mask removal on the planning data including mask content based on the mask mapping information it stores to obtain the planning data without mask, and then the planning executor processes the planning data without mask.

[0145] Among them, mask removal means replacing the target character in the planning data with the corresponding parameter name and / or function name in the mask mapping information.

[0146] In one embodiment, the target function is a function version among all function versions corresponding to a first basic function, and the first basic function is one of the multiple basic functions.

[0147] In the present invention, the difference in function versions can be understood as the difference in function codes.

[0148] For example, if the first basic function is set as an image denoising function, then all function versions corresponding to the first basic function may include: Version V1 (encoded using the idea of image denoising algorithm a) and Version V2 (encoded using the idea of image denoising algorithm b with better image denoising effect).

[0149] In this embodiment, for the case where there may be multiple function versions of a basic function, by configuring / building corresponding function information for each function version of each basic function, not only can the risk that the function codes of different function versions of the same basic function are mixed, resulting in abnormal code reading and further making the basic function unavailable be avoided; but also the situation where a user misinterprets the function of one function version of a basic function as the function of another function version of the same basic function can be avoided, which can further improve the information management effect of multiple basic functions.

[0150] In one embodiment, step S2 includes:

[0151] Step S21: Concatenate the first text, the second text, and the third text to obtain the function information of the target function.

[0152] Exemplarily, in the above concatenation process, a specific delimiter (such as "-") can be used to separate the first text, the second text, and the third text, so that the first text / second text / third text can be accurately and completely read out from the function information subsequently.

[0153] In this embodiment, the method of text concatenation is applied to avoid information loss during the formation of function information and ensure the complete storage of the function code of the target function and its related information in the function information.

[0154] Among them, the related information may include the second text and the third text.

[0155] In one embodiment, after step S2, the method further includes:

[0156] Step S3: Receive an information update instruction for updating the function information of the target function;

[0157] Step S4: In response to the information update instruction, update the function information of the target function to obtain the updated information of the target function;

[0158] Among them, the second text in the updated information is different from the second text in the function information, and / or the third text in the updated information is different from the third text in the function information.

[0159] In this embodiment, through the above settings, the update of function information is supported to facilitate the user to correct / supplement and improve the content of the second text / third text in the function information.

[0160] For example, in order to improve the feasibility of the task planning output by the task planning large model, when optimizing the text format of the second text segment, the second text in the function information can be understood as the function description before optimization, and the second text in the update information can be understood as the function description after optimization.

[0161] Please refer to Figure 4 , Figure 4 which is a schematic structural diagram of an information generation device based on the computing power of an intelligent computing center provided by the present invention. As Figure 4 shown, the information generation device 400 based on the computing power of the intelligent computing center includes:

[0162] A text acquisition module 401, configured to acquire a first text, a second text, and a third text of a target function, where the target function is one of a plurality of basic functions for implementing general functions in a target service, the target service includes a visual processing service based on a computer vision agent, the first text is used to represent the function code of the target function, the second text is used to represent the function function of the target function, the third text is used to represent the application information of the target function, and the application information of the target function includes: the usage log of the target function and the function encapsulation information of the target function, and the usage log of the target function includes service data processed by the target function;

[0163] An information generation module 402, configured to fuse the first text, the second text, and the third text to obtain function information of the target function, where the function information is used to be loaded by a function manager to import and run the function code of the target function.

[0164] In one embodiment, the second text includes at least one of the following:

[0165] A first text segment, and the version type of the first text segment is a natural language version;

[0166] A second text segment, and the version type of the second text segment is a function mode version in an unmasked state;

[0167] A third text segment, and the version type of the third text segment is a function mode version in a masked state;

[0168] wherein, the first text segment, the second text segment, and the third text segment are all used to represent the function function of the target function.

[0169] In one embodiment, when the second text includes the third text segment, the function manager stores mask mapping information for restoring the third text segment to the second text segment.

[0170] In one embodiment, the target function is one function version among all function versions corresponding to a first base function, and the first base function is one of the multiple base functions.

[0171] In one embodiment, the information generation module 402 includes:

[0172] A text splicing unit for splicing the first text, the second text, and the third text to obtain function information of the target function.

[0173] In one embodiment, the information generation device 400 based on the computing power of the intelligent computing center further includes:

[0174] An instruction receiving module for receiving an information update instruction for updating the function information of the target function;

[0175] An information update module for updating the function information of the target function in response to the information update instruction to obtain updated information of the target function;

[0176] Wherein, the second text in the updated information is different from the second text in the function information, and / or the third text in the updated information is different from the third text in the function information.

[0177] The information generation device based on the computing power of the intelligent computing center provided by the present invention can implement each process of the above-mentioned information generation method based on 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 are not elaborated here.

[0178] It should be noted that the information generation device based on the computing power of the intelligent computing center in the present invention can be a device, or a component, an integrated circuit, or a chip in an electronic device.

[0179] 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 stored in the memory 501 and running on the processor 502. When the program or instruction is executed by the processor 502, it can implement Figure 1 any step in the corresponding information generation method embodiment based on the computing power of the intelligent computing center and achieve the same beneficial effects, which are not elaborated here.

[0180] Among them, the processor 502 can be a CPU, an ASIC, an FPGA, or a GPU.

[0181] Those of ordinary skill in the art can understand that all or part of the steps of implementing the above-mentioned information generation method embodiments based on 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.

[0182] 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 any of the steps in the above-mentioned Figure 3 corresponding information generation method embodiments based on the computing power of the intelligent computing center, and can achieve the same technical effects. To avoid repetition, it 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.

[0183] The terms "first", "second", etc. in the present invention are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. In addition, the terms "comprising" and "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 does not have to be 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, in the present invention, "and / or" is used to represent at least one of the connected objects. For example, A and / or B and / or C represents 7 cases including 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.

[0184] It should be noted that in this article, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not explicitly listed, or also includes elements inherent to this process, method, article or device. Without further limitation, an element defined by the statement "comprising one..." does not exclude the existence of another identical element in the process, method, article or device comprising that element.

[0185] Through the description of the above embodiments, those skilled in the art can clearly understand that the above-described example 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 disk) 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.

[0186] The embodiments of the present invention are described above in conjunction with the accompanying drawings. However, 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.

Claims

1. An information generation method based on the computing power of an intelligent computing center, characterized in that, The method includes: Step S1, obtaining a first text, a second text, and a third text of a target function, where the target function is one of a plurality of basic functions for implementing general functions in a target service, the target service includes a vision processing service based on a computer vision agent, the first text is used to represent the function code of the target function, the second text is used to represent the function function of the target function, the third text is used to represent the application information of the target function, and the application information of the target function includes: the usage log of the target function and the function encapsulation information of the target function, and the usage log of the target function includes service data processed by the target function; Step S2, fusing the first text, the second text, and the third text to obtain function information of the target function, where the function information is used to be loaded by a function manager to import and run the function code of the target function.

2. The method according to claim 1, wherein The second text includes at least one of the following: A first text segment, where the version type of the first text segment is a natural language version; A second text segment, where the version type of the second text segment is a function mode version in an unmasked state; A third text segment, where the version type of the third text segment is a function mode version in a masked state; wherein the first text segment, the second text segment, and the third text segment are all used to represent the function function of the target function.

3. The method according to claim 2, wherein When the second text includes the third text segment, the function manager stores mask mapping information, and the mask mapping information is used to restore the third text segment to the second text segment.

4. The method according to claim 2, wherein The target function is one of all function versions corresponding to a first basic function, and the first basic function is one of the plurality of basic functions.

5. The method according to claim 1, wherein The step S2 includes: Step S21, concatenating the first text, the second text, and the third text to obtain function information of the target function.

6. The method according to claim 5, characterized in that After the step S2, the method further includes: Step S3, receiving an information update instruction for updating the function information of the target function; Step S4, in response to the information update instruction, updating the function information of the target function to obtain updated information of the target function; wherein the second text in the updated information is different from the second text in the function information, and / or the third text in the updated information is different from the third text in the function information.

7. An information generation device based on the computing power of an intelligent computing center, characterized in that, including: A text acquisition module, configured to acquire a first text, a second text, and a third text of a target function, where the target function is one of a plurality of basic functions, and the plurality of basic functions are used to implement general functions in a target service. The target service includes a vision processing service based on a computer vision agent. The first text is used to represent the function code of the target function, the second text is used to represent the function function of the target function, and the third text is used to represent the application information of the target function. The application information of the target function includes: the usage log of the target function and the function encapsulation information of the target function. The usage log of the target function includes service data processed by the target function; An information generation module, configured to fuse the first text, the second text, and the third text to obtain function information of the target function, where the function information is used to be loaded by a function manager to import and run the function code of the target function.

8. A server, characterized in that, Comprising: A processor, a memory, and a program stored on the memory and executable on the processor. When the program is executed by the processor, the steps of the information generation method based on the computing power of an intelligent computing center as described in any one of claims 1 to 6 are implemented.

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 a processor, the steps of the information generation method based on the computing power of an intelligent computing center as described in any one of claims 1 to 6 are implemented.

10. A computer program product, characterized in that, Including computer instructions, which when executed by a processor, implement the steps of the information generation method based on the computing power generation of an intelligent computing center as described in any one of claims 1 to 6.

Images