Software product introduction content generation method based on large model and related equipment

Abstract

The embodiment of the invention provides a software product introduction content generation method based on a large model and related equipment, and belongs to the technical field of artificial intelligence. The method comprises the steps that reference data generated by software product introduction content is obtained, the reference data represents a case and a template associated with a software product, target content extraction is conducted on the reference data, large model prompt content is obtained, then an access link of the software product is accessed through a trained first large model, interaction operation is conducted on the software product, and the software product is obtained. And through the trained second large model, according to the product operation video and the large model prompt content, generating software product introduction content in a video form. According to the scheme, the operation video of the software product is intelligently extracted through the first large model, the software product introduction content including the video form is intelligently generated by combining the second large model with the large model prompt content of the related cases and templates, and the generation efficiency and the display effect of the software product introduction content are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of artificial intelligence, in particular to a software product introduction content generation method based on a large model and related equipment. BACKGROUND

[0002] At present, the company usually manually writes the introduction articles such as software product description and operation according to the product situation by relevant personnel. In order to improve the output efficiency of the company articles, the related technology usually uses a large language model as an auxiliary tool to generate article content. This requires the user to input a large number of software product related descriptions and prompt words into the large language model, and then the large language model polishes the user input to output the software product introduction article. This way is still low in efficiency for the user. In addition, with the development of multimedia technology, a single software product introduction article cannot intuitively display the software product, and the software product introduction effect is poor. SUMMARY

[0003] The main purpose of the embodiments of the present application is to provide a software product introduction content generation method based on a large model and related equipment, which aims to improve the generation efficiency and display effect of software product introduction content.

[0004] To achieve the above purpose, one aspect of the embodiments of the present application provides a software product introduction content generation method based on a large model, comprising the following steps: Obtain reference data for generating software product introduction content, wherein the reference data represents cases and templates associated with the software product; Extract target content from the reference data to obtain large model prompt content; Access the access link of the software product through the trained first large model, and perform interactive operation on the software product to obtain product operation video; Generate software product introduction content in the form of video through the trained second large model according to the product operation video and the large model prompt content.

[0005] In some embodiments, the second large model is trained by the following steps: According to the set rules, collect reference data, product operation video and real introduction content of different software products; Respectively extract target content from the reference data of different software products to obtain corresponding large model prompt content; Generate training data set according to the large model prompt content, product operation video and real introduction content of different software products; Fine tune the pre-trained second large model according to the training data set to obtain the trained second large model.

[0006] In some embodiments, the reference data includes text introduction data of cases and templates, website links of cases and templates, the target content extraction on the reference data obtains large model prompt content, including the following steps: Perform named entity recognition on the text introduction data through a knowledge extraction model to obtain first prompt content; Obtain second prompt content by accessing the website links and extracting the content in the website links; According to the first prompt content and the second prompt content, obtain large model prompt content.

[0007] In some embodiments, the generation of software product introduction content in the form of a video based on the product operation video and the large model prompt content through the trained second large model includes the following steps: According to the first prompt content in the large model prompt content, determine the industry and business classification of the software product; According to the industry and business classification, select the expert large model of the corresponding field in the second large model; Generate software product introduction content through the expert large model based on the second prompt content and the product operation video.

[0008] In some embodiments, the large model-based software product introduction content generation method further includes the following steps: Obtain feedback evaluation information of the user on the software product introduction content; Iteratively optimize the second large model according to the feedback evaluation information.

[0009] In some embodiments, the access of the access link of the software product through the trained first large model and the interactive operation on the software product to obtain a product operation video include the following steps: Access the access link of the software product through the first large model to obtain an interactive page of the software product; According to the specified expected function, identify the target component in the interactive page and operate the target component to obtain a product operation video.

[0010] To achieve the above-mentioned purpose, another aspect of the embodiment of the present application proposes a large model-based software product introduction content generation system, including: The first module is configured to obtain reference data for software product introduction content generation, and the reference data represents cases and templates associated with the software product; The second module is configured to perform target content extraction on the reference data to obtain large model prompt content; The third module is configured to access an access link of the software product by using the trained first large model, and to interact with the software product to obtain a product operation video. The fourth module is configured to generate software product introduction content in the form of a video according to the product operation video and the large model prompt content by using the trained second large model.

[0011] To achieve the above object, another aspect of the embodiment of the present application provides an electronic device, which comprises a memory and a processor, the memory stores a computer program, and the processor implements the above method when executing the computer program.

[0012] To achieve the above object, another aspect of the embodiment of the present application provides a computer readable storage medium, which stores a computer program, and the computer program implements the above method when executed by a processor.

[0013] To achieve the above object, another aspect of the embodiment of the present application provides a computer program product, which comprises a computer program, and the computer program implements the above method when executed by a processor.

[0014] The embodiment of the present application at least has the following beneficial effects: the present application provides a software product introduction content generation method, system, electronic device, storage medium and program product based on a large model, the scheme obtains reference data for generating software product introduction content, the reference data represents cases and templates associated with the software product, target content is extracted from the reference data to obtain large model prompt content, then a first trained large model is used to access an access link of the software product and to interact with the software product to obtain a product operation video, and then a second trained large model is used to generate software product introduction content in the form of a video according to the product operation video and the large model prompt content. The scheme intelligently extracts the operation video of the software product by using the first large model, and intelligently generates software product introduction content in the form of a video by using the second large model in combination with the large model prompt content of the related cases and templates, thereby improving the generation efficiency and display effect of the software product introduction content. BRIEF DESCRIPTION OF DRAWINGS

[0015] Figure 1 is a flowchart of the software product introduction content generation method based on a large model provided by the embodiment of the present application; Figure 2 is a hardware structure schematic diagram of the electronic device provided by the embodiment of the present application. DETAILED DESCRIPTION

[0016] In order to make the purposes, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and not intended to limit the present application. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with embodiments of the present application. They are only examples of devices and methods consistent with some aspects of the embodiments of the present application as detailed in the appended claims.

[0017] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs. The terms used herein are only for the purpose of describing the embodiments of the present application and are not intended to limit the present application.

[0018] Before the embodiments of the present application are described in detail, first, some nouns and terms involved in the embodiments of the present application are described, and the nouns and terms involved in the embodiments of the present application are applicable to the following explanations.

[0019] NER (Named Entity Recognition, named entity recognition) is a basic technology in the field of natural language processing. Its core task is to automatically identify entities with specific meanings from unstructured text and classify them according to predefined categories.

[0020] Large model refers to a deep learning model with large-scale parameters (usually reaching tens of billions or even trillions) and complex computing structure. It is trained through massive data and can understand, generate and process natural language, images, audio and other types of information. The core features of large models include strong generalization ability (can adapt to different tasks), high universality (not limited to specific fields) and emergence (may produce unexpected new abilities), and their technical foundation mainly relies on the Transformer architecture, which is a key driving force for the development of artificial intelligence technology.

[0021] SEO (Video Search Engine Optimization) is a systematic process of optimizing video content, titles, descriptions, tags, thumbnails and other elements, and combining technical means and promotion strategies to improve the visibility and ranking of videos in search engine result pages (such as Google, YouTube). The core goal is to make video content more easily understood and indexed by search engines, so that it can be displayed first when users search for relevant keywords, attract more precise traffic and improve user engagement.

[0022] In the related art, a large language model is usually used as an auxiliary tool to generate article content, which requires the user to input a large number of software product related instructions and prompt words to the large language model, and then the large language model polishes the user input to output a software product introduction article. This way is still low in efficiency for the user. In addition, with the development of multimedia technology, a single software product introduction article cannot intuitively display the software product, and the software product introduction effect is poor.

[0023] Therefore, the embodiments of the present application provide a software product introduction content generation method based on a large model and related equipment. The scheme intelligently extracts an operation video of a software product through a first large model, and intelligently generates software product introduction content containing a video form by combining a related case and a template of a large model prompt content through a second large model, thereby improving the generation efficiency and display effect of the software product introduction content.

[0024] The software product introduction content generation method based on a large model provided by the embodiments of the present application relates to the field of artificial intelligence. The software product introduction content generation method based on a large model provided by the embodiments of the present application can be applied to a terminal, can be applied to a server, and can also be software running in a terminal or a server. In some embodiments, the terminal can be a smart phone, a tablet computer, a notebook computer, a desktop computer, a smart speaker, a smart watch, a vehicle-mounted terminal, and the like, but is not limited thereto. The server end can be configured as an independent physical server, can be configured as a server cluster or a distributed system composed of multiple physical servers, can be configured as a cloud server providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDNs, and basic cloud computing services such as big data and artificial intelligence platforms, and the server can also be a node server in a blockchain network. The software can be an application that implements the software product introduction content generation method based on a large model, and the like, but is not limited to the above forms.

[0025] The application is operable in a variety of general purpose or special purpose computer systems environments or configurations. Examples of well known computing systems, environments, and / or configurations that can be suitable for use with the application include personal computers, server computers, handheld or laptop devices, tablet devices, multiprocessor systems, microprocessor-based systems, set top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments that include any of the above systems or devices, and the like. The application can be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, and the like, that perform particular tasks or implement particular abstract data types. The application can also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules can be located in local and remote computer storage media including memory storage devices.

[0026] Figure 1 is an optional flowchart of the method for generating software product introduction content based on a large model provided by the embodiments of the application, Figure 1 The method in the above embodiment can include, but is not limited to, steps S101 to S104.

[0027] S101, obtaining reference data for generating software product introduction content, the reference data representing cases and templates associated with the software product; S102, extracting target content from the reference data to obtain large model prompt content; S103, accessing the access link of the software product through the trained first large model, and performing interactive operation on the software product to obtain product operation video; S104, generating software product introduction content in the form of video through the trained second large model according to the product operation video and the large model prompt content.

[0028] The steps S101 to S104 shown in the embodiments of the application first obtain reference data for generating software product introduction content, the reference data representing cases and templates associated with the software product, extract target content from the reference data to obtain large model prompt content, then access the access link of the software product through the trained first large model, and perform interactive operation on the software product to obtain product operation video, and finally generate software product introduction content in the form of video through the trained second large model according to the product operation video and the large model prompt content. The embodiments of the application intelligently extract the operation video of the software product through the first large model, and intelligently generate software product introduction content in the form of video by combining the large model prompt content of the related cases and templates through the second large model, thereby improving the generation efficiency and display effect of the software product introduction content.

[0029] In step S101 of some embodiments, the case associated with the software product to be introduced in the present embodiment can refer to the introduction case (such as an introduction article, a video or a picture, etc.) of other software products similar to the software product in industry, business or function, and the case refers to the specific introduction content published on the Internet about the associated software product, including but not limited to video, text and picture forms, etc. The template associated with the software product to be introduced in the present embodiment refers to the reference template stored in the local database, including but not limited to video, text and picture forms, etc., which can be obtained by parsing according to the business data structure from the database. Illustratively, the user first determines the software product to be introduced, for example, a certain information management platform, and then collects the cases and templates associated with the information management platform as reference data. The collected reference data specifically includes the text introduction data and website links of the cases or templates, and the text introduction data includes the name of the case and the template, the business to which the case and the template belong, the industry classification of the case and the template, etc. The relevant database or network can be accessed to collect data through the set association rules (such as software products in a certain industry or a certain business) and collection items (such as the case or the name of the case, the industry attribute, the business attribute, the text, etc. Related content), and then the collected data is preprocessed by data extraction and cleaning, etc. to obtain standardized reference data. It can be understood that the association rules and the collection items can be defined and adjusted according to actual needs.

[0030] In step S102 of some embodiments, in order to quickly locate the key content in the reference data and form the input format of the second large model, the target content extraction is performed on the reference data to obtain the large model prompt content. Specifically, the text introduction data can be subjected to business, industry, function, etc. named entity extraction to obtain the first prompt content, and the access links of the cases and templates are accessed to extract the specific introduction cases and specific introduction templates of the associated products to obtain the second prompt content. The first prompt content and the second prompt content are integrated according to the set format rule to obtain the large model prompt content.

[0031] In step S103 of some embodiments, the trained first large model is a large model that fine-tunes a general large model and integrates automatic operation and video generation functions. The first large model can intelligently identify the interactive pages of a software product and automatically interact with the software product according to user instructions to generate a product operation video. Specifically, the first large model includes a page recognition module, an operation decision module, an operation execution module, and a video sampling module. The first large model starts the video sampling module when entering the software product according to the access link of the software product, and records the video at a certain frame rate; identifies all page contents of the current page through the page recognition module; intelligently decides the page contents through the operation decision module based on the page recognition module, determines the next operation, executes the next operation output by the operation decision module through the operation execution module, and enters the next page. The page recognition module, the operation decision module, and the operation execution module are jointly trained.

[0032] In step S104 of some embodiments, the product operation video obtained above and the large model prompt content are input into the trained second large model to output software product introduction content in the form of a video. Specifically, the software product introduction content output by the first large model includes but is not limited to introduction videos and articles of the software product and their SEO information, industry classification, business classification, etc. The second large model of the present embodiment can select a large language model that supports Chinese well and has excellent instruction compliance, such as GLM4.

[0033] According to some embodiments of the present application, step S102 can include but is not limited to the following steps: S201, performing named entity recognition on the text introduction data by a knowledge extraction model to obtain first prompt content; S202, obtaining second prompt content by accessing a website link and extracting content in the website link; S203, obtaining large model prompt content according to the first prompt content and the second prompt content.

[0034] In step S201 of some embodiments, the text introduction data includes the names of cases and templates, the business to which the cases and templates belong, the industry classification of the cases and templates, and other text contents. For the text introduction data, a general large language model can be used first to extract keywords and summarize with prompt words engineering; then a knowledge extraction model is used to perform named entity recognition on the output of the general large language model to obtain a named entity list about the text introduction data, which can effectively eliminate irrelevant information such as brand words, adjectives, and quantifiers, and further extract and summarize the named entity list by the general model to obtain the first prompt content. The knowledge extraction model of the present embodiment can select a model that fits the business field scene and has excellent recognition accuracy, such as a knowledge extraction model in the field of software development.

[0035] In step S202 of some embodiments, by accessing the website link of the case and the template, the specific content introducing the case and the template in the website link is extracted, which can be in the format of text, picture and video, etc. Then, the multi-modal large model is used to recognize and summarize the extracted specific content to obtain the second prompt content. The multi-modal large model of the present embodiment can select a large language model with better recognition accuracy and better summarization and description capability, such as GLM4-V.

[0036] In step S203 of some embodiments, the first prompt content and the second prompt content are integrated according to the set format rule to obtain the large model prompt content. In the present embodiment, the named entity recognition of the text introduction data is performed by the knowledge extraction model, which can extract the key content in the text. The specific content of the case and the template is automatically accessed, and the multi-model large model is used to summarize the specific content, so that the reference data of the case and the template is processed into more concise large model prompt content, which improves the efficiency and effect of the subsequent second large model generating product introduction content.

[0037] According to some embodiments of the present application, step S104 can include but is not limited to the following steps: S301, determining the industry and business classification of the software product according to the first prompt content in the large model prompt content; S302, selecting an expert large model in the second large model in the corresponding field according to the industry and business classification; S303, generating software product introduction content by the expert large model according to the second prompt content and the product operation video.

[0038] In the present embodiment, the second large model includes a plurality of expert large models applied to different industry and business combinations. The expert large model is a large language model specially used to generate product introduction text and video content in the corresponding industry and business. Each expert large model is obtained by fine-tuning a pre-trained general large language model with corresponding industry and business data, and the fine-tuning process can be realized by Lora fine-tuning technology. In the second large model, first, the industry and business classification of the software product is determined according to the first prompt content in the large model prompt content, then the expert large model in the corresponding field of the second large model is selected according to the industry and business classification, and finally the second prompt content and the product operation video are input into the selected expert large model to create product introduction content, thereby obtaining software product introduction content. In the present embodiment, the appropriate expert large model is selected based on the first prompt content to create content, which improves the style diversity and.

[0039] According to some embodiments of the present application, the second large model in step S104 is trained by the following steps: S401, collect reference data, product operation videos and real introduction contents of different software products according to a set rule; S402, extract target content from the reference data of different software products respectively to obtain corresponding large model prompt content; S403, generate a training data set according to the large model prompt content, product operation videos and real introduction contents of different software products; S404, fine-tune the pre-trained second large model according to the training data set to obtain a trained second large model.

[0040] In step S401 of some embodiments, the collection method of reference data and product operation videos of different software products in the training process is the same as that of the previous embodiments, which will not be repeated here. The real introduction content refers to the introduction content of the corresponding software product created by artificial participation, which can refer to the software product introduction content created completely by artificial participation, or the software product introduction content created by a large language model and adjusted by artificial participation.

[0041] In step S402 of some embodiments, the target content of the reference data of different software products is extracted to obtain corresponding large model prompt content. The target content extraction of the reference data in this embodiment is the same as that of the above embodiments, which will not be repeated here.

[0042] In step S403 of some embodiments, a training data set is generated according to the large model prompt content, product operation videos and real introduction contents of different software products, wherein the training data set includes a plurality of sample data, each sample data includes the large model prompt content, product operation videos and real introduction contents of a certain software product, the large model prompt content and product operation videos are input data of the second large model, and the real introduction content is a sample label.

[0043] In step S404 of some embodiments, the input data in the training data set is input into the second large model for forward propagation to obtain industry and business classification and product introduction content, and then the second large model is adjusted according to the real introduction content adjusted by artificial adjustment to obtain a trained second large model. Specifically, the second large model includes an enhancement module and a plurality of expert large models, and the enhancement module is used to determine the corresponding industry and business classification according to the input first prompt content. In the training process of the second large model, the enhancement model in the second large model is adjusted according to the actual industry and business classification indicated in the real introduction content. After adjusting the enhancement module, the loss value of the selected expert large model in the creation process is calculated according to the difference between the product introduction content output by the second large model and the real introduction content, and the expert large model is updated in reverse according to the loss value, so that the second large model can be adapted to the creation of software products in various industries and businesses.

[0044] According to some embodiments of the present application, the software product introduction content generation method based on a large model according to the embodiments of the present application can further include but is not limited to the following steps: S501, obtaining feedback evaluation information of the user on the software product introduction content; S502, iteratively optimizing the second large model according to the feedback evaluation information.

[0045] In the present embodiment, in the actual creation process, after inputting the product operation video of the software product to be introduced and the large model prompt content into the fine-tuned second large model, the software product introduction content in the form of video is obtained. The software product introduction content output by the second large model is audited by the auditing personnel, and the software product introduction content that passes the audit is published; the feedback evaluation information of the software product introduction content that does not pass the audit is obtained to iteratively optimize the second large model. Specifically, the software product introduction content that does not pass the audit and the feedback evaluation information thereof can be arranged into a new training data set, and the large model is fine-tuned and trained regularly to achieve the purpose of continuous iterative optimization.

[0046] According to some embodiments of the present application, step S103 can include but is not limited to the following steps: S601, accessing the access link of the software product by the first large model to obtain the interactive page of the software product; S602, identifying the target component in the interactive page according to the specified expected function, and operating the target component to obtain the product operation video.

[0047] In the embodiment, the first large model is improved on the basis of a general language model. Specifically, the general large language model is configured with a video recording function, a page collection and operation execution function, and the large language model is trained to be capable of recognizing the content in the page and making intelligent decisions according to the recognized content, so that the first large model can automatically operate the software product according to the user instruction and output the operation process video of the software product. Exemplarily, a user needs to obtain the operation process of the shopping function of an e-commerce software, and can input the access link of the e-commerce software and specify the expected operation function as the shopping function operation to the first large model. The first large model accesses the corresponding e-commerce software according to the access link, and simultaneously starts the video sampling model to record the video; all page contents of the current page are recognized by the page recognition module, which can be a convolutional neural network for image recognition, and the page contents include page title, component position and component function in the page, and the like; the operation decision module makes intelligent decisions according to the page contents recognized by the page recognition module, determines the optimal operation action in the next step, so that the final series of actions can realize the shopping function, and the operation decision module can be a rule-based or probabilistic decision model; the operation execution module executes the next operation output by the operation decision module, enters the next page, and thus obtains the shopping function operation process video. The embodiment automatically operates the software product by the first large model and outputs the product operation video, and then the product operation video is input into the second large model to generate the product introduction video article (such as operation tutorial), thereby improving the generation efficiency of the product introduction video article.

[0048] The embodiment of the present application also provides a software product introduction content generation system based on a large model, comprising: A first module configured to obtain reference data for software product introduction content generation, the reference data representing cases and templates associated with the software product; A second module configured to extract target content from the reference data to obtain large model prompt content; A third module configured to access an access link of the software product by the trained first large model, and interact with the software product to obtain a product operation video; A fourth module configured to generate software product introduction content in the form of a video based on the product operation video and the large model prompt content by the trained second large model.

[0049] It can be understood that the method content in the above method embodiments is applicable to the system embodiment, the system embodiment specifically implements the same functions as the above method embodiments, and achieves the same beneficial effects as the above method embodiments.

[0050] The embodiment of the present application further provides an electronic device, which comprises a memory and a processor. The memory stores a computer program, and the processor executes the computer program to realize the method described above. The electronic device can be any intelligent terminal, such as a tablet computer or a vehicle-mounted computer.

[0051] It can be understood that the contents in the method embodiments described above are applicable to the device embodiments, the device embodiments specifically realize the functions of the method embodiments, and achieve the same beneficial effects as the method embodiments.

[0052] Please refer to Figure 2 , Figure 2 The hardware structure of the electronic device of another embodiment is illustrated, and the electronic device comprises: The processor 901 can be implemented in a general-purpose CPU (Central Processing Unit), a microprocessor, an ASIC (Application Specific Integrated Circuit), or one or more integrated circuits, and is used to execute related programs to realize the technical solutions provided by the embodiments of the present application. The memory 902 can be implemented in the form of a ROM (Read Only Memory), a static storage device, a dynamic storage device, or a RAM (Random Access Memory). The memory 902 can store an operating system and other application programs. When the technical solutions provided by the embodiments of the present application are implemented by software or firmware, the related program codes are stored in the memory 902 and are called and executed by the processor 901 to realize the method of the embodiments of the present application. The input / output interface 903 is used to realize information input and output. The communication interface 904 is used to realize the communication interaction between the device and other devices. The communication can be realized by a wired manner (for example, a USB, a network cable, etc.) or a wireless manner (for example, a mobile network, WIFI, Bluetooth, etc.). The bus 905 is used to transmit information between various components (for example, the processor 901, the memory 902, the input / output interface 903, and the communication interface 904) of the device. The processor 901, the memory 902, the input / output interface 903, and the communication interface 904 are connected to each other through the bus 905 to realize the communication connection between them in the device.

[0053] The embodiment of the present application further provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to realize the method.

[0054] It can be understood that the contents in the method embodiments are applicable to the storage medium embodiments, the storage medium embodiments specifically realize the functions of the method embodiments, and achieve the same beneficial effects as the method embodiments.

[0055] The embodiment of the present application further provides a computer program product, which comprises a computer program, and the computer program is executed by a processor to realize the method.

[0056] It can be understood that the contents in the method embodiments are applicable to the program product embodiments, the program product embodiments specifically realize the functions of the method embodiments, and achieve the same beneficial effects as the method embodiments.

[0057] The memory is a non-transitory computer readable storage medium, and can be used to store a non-transitory software program and a non-transitory computer executable program. In addition, the memory can include a high-speed random access memory, and can also include a non-transitory memory, for example, at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state memory device. In some embodiments, the memory can optionally include a memory remotely arranged relative to the processor, and the remote memory can be connected to the processor through a network. Examples of the network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0058] The embodiments described in the embodiments of the present application are used to more clearly illustrate the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application. Those skilled in the art can know that, with the evolution of technology and the appearance of new application scenarios, the technical solutions provided by the embodiments of the present application are also applicable to similar technical problems.

[0059] Those skilled in the art can understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and can include more or fewer steps than the figures, or combine certain steps, or different steps.

[0060] The system embodiments described above are only schematic, and the modules described as separate components can or can not be physically separate, that is, can be located in one place, or can be distributed on multiple network units. According to actual needs, part or all of the modules can be selected to realize the purposes of the embodiments.

[0061] Those skilled in the art can understand that all or some of the steps in the method disclosed above, the function modules / units in the system and the device can be implemented as software, firmware, hardware or appropriate combination thereof.

[0062] The terms "first", "second", "third", "fourth" and the like in the description of the application and in the claims, if any, are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the use of these terms herein is to be construed to cover a general order and / or structure unless otherwise indicated. Furthermore, the terms "comprise", "have" and any variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises, has or includes a list of elements is not necessarily limited to those elements, but can include other elements not expressly listed or inherent to such process, method, article, or apparatus.

[0063] It should be understood that, in the present application, "at least one" means one or more, and "multiple" means two or more. "And / or" is used to describe the relationship between associated objects, which means that there can be three relationships, for example, "A and / or B" can mean that there are three cases: only A, only B, and A and B at the same time, where A and B can be singular or plural. The character " / " generally represents an "or" relationship between the associated objects. "At least one of the following" or similar expressions means any combination of these items, including any combination of single or multiple items. For example, at least one of a, b or c can mean a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0064] In several embodiments provided in the present application, it should be understood that the disclosed system and method can be implemented in other ways. For example, the above-described system embodiments are only illustrative, for example, the division of the above-mentioned modules is only a logical function division, and actual implementation can have another division manner, for example, multiple modules can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed each other can be indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0065] The modules described above as separate components may or may not be physically separate, and the components shown as modules may or may not be physical modules, i.e., may be located in one place, or may be distributed to multiple network elements. Some or all of the elements can be selected according to actual needs to achieve the purpose of the embodiment.

[0066] In addition, the functional modules in each embodiment of the present application can be integrated in one processing module, or each module can exist physically alone, or two or more modules can be integrated in one module. The integrated module can be realized in the form of hardware or in the form of a software functional module.

[0067] The integrated module, if realized in the form of a software functional module and sold or used as an independent product, can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application, essentially or the part that contributes to the prior art, or all or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes multiple instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods of the embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various program storage media.

[0068] The preferred embodiments of the embodiments of the present application are described above with reference to the accompanying drawings, and the scope of the rights of the embodiments of the present application is not limited thereto. Any modifications, equivalent replacements and improvements made by those skilled in the art without departing from the scope and essence of the embodiments of the present application shall be within the scope of the rights of the embodiments of the present application.

Claims

1. A method for generating software product introduction content based on a large model, characterized in that, Includes the following steps: The reference data generated from the software product introduction content is obtained, and the reference data represents cases and templates associated with the software product; The target content is extracted from the reference data to obtain the large model prompt content; By accessing the access link of the software product through the first trained model and interacting with the software product, a product operation video is obtained. Using the trained second large model, software product introduction content in video format is generated based on the product operation video and the prompts from the large model.

2. The method according to claim 1, characterized in that, The second major model is obtained through the following steps: Collect reference data, product operation videos, and authentic introductions of different software products according to the set rules; The target content is extracted from the reference data of different software products to obtain the corresponding large model prompt content; Training datasets are generated based on large-scale model prompts, product operation videos, and real-world introductions for different software products. The pre-trained second-largest model is fine-tuned based on the training dataset to obtain the trained second-largest model.

3. The method according to claim 1, characterized in that, The reference data includes textual descriptions of cases and templates, and website links to cases and templates. Extracting target content from the reference data to obtain large-scale model prompts includes the following steps: The first prompt content is obtained by performing named entity recognition on the text description data using a knowledge extraction model. The second prompt content is obtained by accessing the website link and extracting the content from the website link; Based on the first and second prompts, the prompts for the large model are obtained.

4. The method according to claim 3, characterized in that, The process of generating software product introduction content in video format using the trained second large model, based on the product operation video and the prompts from the large model, includes the following steps: Based on the first prompt in the large model prompt, determine the industry and business classification of the software product; Select the expert model for the corresponding field in the second major model based on the industry and business classifications described above; Based on the second prompt and the product operation video, the expert model generates software product introduction content.

5. The method according to claim 2, characterized in that, The method for generating software product introduction content based on a large model also includes the following steps: Obtain user feedback and evaluation information regarding the software product introduction content; The second major model is iteratively optimized based on the feedback evaluation information.

6. The method according to claim 1, characterized in that, The process of accessing the software product via the pre-trained first model, interacting with the software product, and obtaining the product operation video includes the following steps: By accessing the access link of the software product through the first large model, the interactive page of the software product can be obtained; Identify the target component in the interactive page according to the specified desired function, and operate the target component to obtain a product operation video.

7. A software product introduction content generation system based on a large model, characterized in that, include: The first module is used to obtain reference data for generating software product introduction content, wherein the reference data represents cases and templates associated with the software product; The second module is used to extract target content from the reference data to obtain the large model prompt content; The third module is used to access the access link of the software product through the trained first model, and to perform interactive operations on the software product to obtain product operation videos. The fourth module is used to generate software product introduction content in video format based on the product operation video and the prompts from the large model, using the trained second large model.

8. An electronic device, characterized in that, The electronic device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method according to any one of claims 1 to 6.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 6.

Images