Intelligent training method and system based on large generative model

Abstract

The invention provides an intelligent training method and system based on a generative large model. The method comprises the following steps: acquiring a training demand of a user; analyzing the training demand, and obtaining a training demand structure and a training demand mode; the training requirement structure comprises organization and arrangement modes of all components in the training process; the training demand mode comprises a specific method or strategy adopted in the training process; and generating an intelligent training service based on the training demand structure and the training demand mode by using a generative large model. According to the invention, one-to-one training service can be provided for the user, targeted and instant training service can be effectively provided around the task target, and the training service which is more standardized and meets the actual application scene is provided while the private requirement of the user is met. Through application of the method and the system, the service capability of individuals can be effectively improved, the working efficiency is improved, enterprises can greatly shorten the competency growth cycle of the post capability of employees, and cost reduction and efficiency improvement are realized.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of business training, and particularly relates to an intelligent training method and system based on a generative large model. BACKGROUND

[0002] Traditional training methods include apprenticeship or collective business training. For apprenticeship, the master has sufficient work experience and can provide instant feedback and correction, but lacks a unified methodology and scientific training, resulting in different masters having different work habits, which makes it difficult to provide standardized employee operations for enterprises. For collective business training, a large number of manpower is needed to prepare training course content, and the trainer is under tremendous pressure and is forced to choose a flowchart training mode, such as the currently widely used "online classroom", which mainly records video courses and requires employees to learn within a specified period. This focuses more on one-way output and is difficult to meet the needs of enterprises and the characteristics of employees' positions. Moreover, there is a lack of quantifiable indicators to evaluate training effectiveness, which greatly increases the cost of enterprise training. SUMMARY

[0003] The purpose of the present application is to provide an intelligent training method and system based on a generative large model, which can improve the pertinence and efficiency of training and reduce the cost of training.

[0004] In a first aspect, the present application provides an intelligent training method based on a generative large model, which comprises:

[0005] Obtaining the training needs of the user;

[0006] Analyzing the training needs to obtain the training needs structure and the training needs mode; the training needs structure includes the organization and arrangement of each component in the training process; the training needs mode includes specific methods or strategies used in the training process;

[0007] Generating intelligent training services based on the training needs structure and the training needs mode using a generative large model.

[0008] In an implementation form of the first aspect, obtaining the training needs of the user comprises:

[0009] Obtaining the scene characteristics of the training needed; the scene characteristics include the first role characteristics of the user, the second role characteristics of the user interacting with, and the task target;

[0010] Generating corresponding matched training topics based on the scene characteristics of different contents; that is, generating training topics based on the first role characteristics of the user, the second role characteristics of the user interacting with, and the task target.

[0011] In an implementation form of the first aspect, the scenario feature of the training demand of the user is a private coaching scenario; the private coaching scenario comprises a private role, a coach role and a private coaching target;

[0012] The generating large model generates intelligent training service based on the training demand structure, specifically comprising:

[0013] The generating large model provides corresponding standard language structure and answer demonstration to the user based on the training theme;

[0014] The generating large model extracts endless scenario questions based on the training theme and preset category and condition; the scenario questions comprise any one or several of task, time, place, character, character relationship, character emotional state and key event;

[0015] The generating large model outputs the scenario questions based on preset normative instruction structure;

[0016] The generating large model identifies the response of the user to the scenario questions, matches the standard language structure, and outputs response evaluation; wherein,

[0017] When the response does not conform to the standard language structure and there is a structural defect, the generating large model triggers a teaching mode, outputs a prompt for guiding the user to continue the complete response content, until the response conforms to the standard language structure;

[0018] When the response conforms to the standard language structure, the generating large model triggers a question mode and outputs the next scenario question.

[0019] In an implementation form of the first aspect, the scenario feature of the training demand of the user is an assistant trainer scenario; the assistant trainer scenario comprises a trainer role, a trainer assistant role and a training target;

[0020] The generating large model generates intelligent training service based on the training demand structure and the training demand mode, specifically comprising:

[0021] The generating large model extracts training key information based on the training demand;

[0022] The generating large model generates first training materials based on the training demand structure and the training key information;

[0023] The generative large model extracts course outline key information based on the training demand.

[0024] In an implementation form of the first aspect, the generative large model providing the intelligent training assistance service to the user in the role of the trainer in the role of the trainer assistant includes the generative large model providing a course outline generation service to the user in the role of the trainer in the role of the trainer assistant, including:

[0025] The generative large model extracts course outline key information based on the training demand.

[0026] The generative large model generates course outline information based on the training demand structure and the course outline key information.

[0027] The generative large model extracts course outline key information based on the training demand.

[0028] The generative large model generates course outline information based on the training demand structure and the course outline key information.

[0029] In an implementation form of the first aspect, the generative large model providing the intelligent training assistance service to the user in the role of the trainer in the role of the trainer assistant includes the generative large model providing a course outline generation service to the user in the role of the trainer in the role of the trainer assistant, including:

[0030] The generative large model extracts course outline key information based on the training demand.

[0031] The generative large model generates course outline information based on the training demand structure and the course outline key information.

[0032] The generative large model extracts course outline key information based on the training demand.

[0033] The generative large model generates course outline information based on the training demand structure and the course outline key information.

[0034] In an implementation form of the first aspect, the providing, by the generative large model in the role of the trainer assistant, the intelligent training assistance service to the user in the role of the trainer comprises: the generative large model providing, in the role of the trainer assistant, a generative content material service to the user in the role of the trainer, including:

[0035] The generative large model extracts material key information based on the training demand; the training demand includes a preset course outline or the training course outline generated by the generative large model;

[0036] The generative large model generates material information based on the training demand structure, the training course outline and the material key information; the material information includes course knowledge points, case stories and tool methods;

[0037] The generative large model extracts endless material related information based on the training demand mode and preset material categories and conditions, and supplements the material information based on the material related information to generate material content;

[0038] The generative large model generates training material conforming to a preset content structure based on the material information and the material content.

[0039] In an implementation form of the first aspect, the providing, by the generative large model in the role of the trainer assistant, the intelligent training assistance service to the user in the role of the trainer comprises: the generative large model providing, in the role of the trainer assistant, a generative course test question service to the user in the role of the trainer, including:

[0040] The generative large model extracts test question key information based on the training demand; the training demand further includes user-entered knowledge points or the course knowledge points generated by the generative large model;

[0041] The generative large model generates test question information based on the training demand structure, the test question key information and the course knowledge points in the material information;

[0042] The generative large model extracts endless test question related information based on the training demand mode and preset test question categories and conditions, and supplements the test question information based on the test question related information to generate test question content;

[0043] The generative large model generates training test questions conforming to a preset content structure and covering the course knowledge points based on the test question information and the test question content.

[0044] In an implementation form of the first aspect, the generating large model providing intelligent training assistance services to the user in the role of the trainer in the role of the trainer assistant includes: the generating large model providing generating training interaction services to the user in the role of the trainer in the role of the trainer assistant, including:

[0045] The generating large model extracts interactive key information based on the training demand; the training demand includes a preset course outline or the training course outline generated by the generating large model;

[0046] The generating large model generates interactive information based on the training demand structure, the training course outline and the interactive key information; the interactive information includes opening interactive, process interactive and closing interactive;

[0047] The generating large model extracts endless interactive related information based on the training demand mode and preset interactive category and condition, and supplements the interactive information based on the interactive related information to generate interactive content;

[0048] The generating large model generates training interaction in line with the preset content structure based on the interactive information and the interactive content.

[0049] In a second aspect, the present application provides an intelligent training system based on a generating large model, which is configured to:

[0050] An acquisition module is configured to acquire training demand of a user;

[0051] An analysis module is configured to analyze the training demand to acquire training demand structure and training demand mode; the training demand structure includes organization and arrangement of each component in the training process; the training demand mode includes specific methods or strategies used in the training process;

[0052] A training module is configured to generate intelligent training services based on the training demand structure and the training demand mode by using the generating large model trained by the method of the second aspect, to realize the method of the first aspect.

[0053] In a third aspect, the present application provides a computer readable storage medium having a computer program stored thereon, which, when executed by an electronic device, realizes the intelligent training method based on the generating large model of the first aspect.

[0054] In a fourth aspect, the present application provides an electronic device, which includes a processor and a memory;

[0055] The memory is configured to store a computer program;

[0056] The processor is configured to execute a computer program stored in the memory, so that the electronic device executes the intelligent training method based on the generative large model.

[0057] As described above, the intelligent training method and system based on the generative large model has the following beneficial effects:

[0058] (1) The present application can provide one-to-one training services for users, effectively providing targeted training services around task goals, and users can receive immediate feedback and correction.

[0059] (2) The present application can meet the privacy needs of users while also meeting the actual application scenarios.

[0060] (4) Personal application of the present application can effectively improve business capabilities and work efficiency, quickly pass the employee training period, reduce the time and labor costs paid by users for business training, provide more standardized training services, achieve more skilled work skills and stronger psychological qualities of employees, and be able to face multiple scenarios of work environment to maximize work value.

[0061] (5) Enterprise application of the present application can fully focus on enterprise business goals for training, solve the actual problem of improving the competence of external interaction of the post, greatly shorten the competence growth cycle of the post ability of employees, reduce the cost of human, material and financial resources in the process of personnel training, and the loss of training results caused by the turnover of enterprise employees, and achieve cost reduction and efficiency increase of enterprises.

[0062] (6) The present application provides training data recording services, which can make users obtain more real, direct and accurate training result data, so that enterprises can make further training decisions based on training data. BRIEF DESCRIPTION OF DRAWINGS

[0063] Figure 1 A scene schematic diagram of an embodiment of the electronic device of the present application is shown.

[0064] Figure 2 A flowchart of an embodiment of the intelligent training method based on the generative large model is shown.

[0065] Figure 3 A flowchart of an embodiment of the intelligent training method based on the generative large model is shown.

[0066] Figure 4 A flowchart of an embodiment of the intelligent training method based on the generative large model is shown.

[0067] Figures 5a to 5c An interface display diagram of the intelligent training assistance provided by the generative large model in a private training scenario is shown.

[0068] Figure 6 A flowchart of the intelligent training method based on the generative large model according to an embodiment of the present application is shown.

[0069] Figures 7a to 7b An interface display diagram of the intelligent training assistance provided by the generative large model in an assistant training scenario is shown.

[0070] Figures 8a to 8c An interface diagram of the training data according to an embodiment of the present application is shown.

[0071] Figure 9 A flowchart of the model training method according to an embodiment of the present application is shown.

[0072] Figure 10 A structural diagram of the intelligent training system based on the generative large model according to an embodiment of the present application is shown.

[0073] Figure 11 A structural diagram of the electronic device according to an embodiment of the present application is shown. DETAILED DESCRIPTION

[0074] The embodiments of the present application will be described in detail hereinafter with specific reference on the attached drawings. It is to be noted that other advantages and effects of the present application will be more clearly understood from the following description. The present application can be embodied in various ways, and thus the description made herein only illustrates the basic idea of the present application. Therefore, it is to be noted that the embodiments described herein and the features thereof can be combined with each other without conflict.

[0075] It is to be noted that the diagrams provided in the following embodiments only schematically illustrate the basic idea of the present application, and thus the diagrams only show the components related to the present application without showing the number, shape and size of the components in actual implementation. The shape, number and ratio of the components in actual implementation can be arbitrarily changed, and the layout of the components can be more complicated.

[0076] The embodiments of the present application provide an intelligent training method based on a generative large model, which can be applied to an intelligent training device based on a generative large model as shown in the following embodiments. Figure 1 The intelligent training device based on a generative large model according to the present application can include a generative large model 12 and a database 13. In the present application, the intelligent training device based on a generative large model can include a training data interface 14, a model training interface 15, a training result interface 16, a training data interface 17, a model training interface 18 and a training result interface 19. Figure 1In the intelligent training device based on a generative large model, when the user's training needs are characterized by a private coaching scenario, the generative large model 12 provides intelligent training services to the user in the role of a coach, thereby achieving the private coaching goal. After determining the private coaching goal, the generative large model 12 generates and displays a corresponding matching training topic (at least one). Based on the user's selection, the generative large model 12 provides the user with corresponding standard language structures and answer examples. Upon receiving the user's start command, the generative large model 12, based on the training topic and preset categories and conditions, captures an infinite number of scenario questions and outputs the scenario questions based on a preset standardized instruction structure. After the user answers the scenario questions, the generative large model 12 identifies the user's response to the scenario questions, matches the standard language structure, and triggers different modes to provide timely feedback and correction to the user's answer. Correspondingly, when the user's training needs are characterized by an assistant trainer scenario, the generative large model 12 provides intelligent training assistance services to the user in the role of a trainer assistant, thereby achieving the training goal. After determining the training objectives, the generative large-scale model 12 extracts key training information based on the training needs, and generates first training materials based on the training needs structure and the key training information. Then, based on the training needs pattern and preset categories and conditions, the generative large-scale model 12 captures an infinite amount of relevant information, and supplements the first training materials with this information to generate second training materials that conform to a preset content structure. Furthermore, Figure 1 The provided intelligent training device based on generative large models can also manage training data. User training data will be stored in database 13 for management purposes.

[0077] The technical solutions in the embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0078] like Figure 2 As shown, in one embodiment, the intelligent training method based on generative large models of this application includes steps S1-S3.

[0079] S1: Obtain users' training needs.

[0080] Specifically, such as Figure 3 As shown, step S1 includes steps S11-S12.

[0081] S11: Obtain the scenario features that need to be trained; the scenario features include the user's first role features, the second role features of the user's interaction, and the task objectives.

[0082] In some embodiments, the scenario feature of the training needs of the user is a private coaching scenario, which includes a private role, a coach role and a private coaching target. That is, the first role feature of the user is a private role, the second role feature interacting with the user is a coach role, and the private coaching target is a target to be achieved by the intelligent training service provided by the coach role to the private role of the user, including experience marketing, resistance resolution, guided parking, persuasion behavior, guide reception, complaint resolution, new product mastery, offline meeting invitation, expression of praise, expression of refusal.

[0083] In other embodiments, the scenario feature of the training needs of the user is an assistant trainer scenario, which includes a trainer role, a trainer assistant role and a training target. That is, the first role feature of the user is a trainer role, the second role feature interacting with the user is a trainer assistant role, and the training target is a target to be achieved by the intelligent training assistance service provided by the trainer assistant role to the user of the trainer role, including providing training services for enterprises.

[0084] S12: generating a corresponding matched training theme based on the scenario features of different contents; that is, generating a training theme based on the first role feature of the user, the second role feature interacting with the user and the task target.

[0085] In some embodiments, the scenario feature of the training needs of the user is a private coaching scenario, and the private coaching target is experience marketing, and the training theme is a key touch task corresponding to the private coaching target of experience marketing, such as inviting customers to participate in experience, inviting customers to experience, arousing user experience empathy, etc. That is, the coach role provides an intelligent training service to the private role to achieve the private coaching target of experience marketing. At this time, under this private coaching target, the training theme of the intelligent training service provided includes inviting customers to participate in experience, inviting customers to experience, arousing user experience empathy, etc.

[0086] It should be noted that when the scenario feature is a private coaching scenario, the key touch task (training theme) corresponding to each private coaching target is based on experience extraction or real situation extraction scenario. That is, under each private coaching target, there should be at least one training theme.

[0087] In some embodiments, the scenario feature of the user's training demand is a private training scenario; the private training scenario includes a private role, a coach role, and a private training goal. Then, generating the intelligent training training service based on the training demand structure by the generative large model includes: the generative large model provides the user in the private role with the intelligent training training service in the coach role to achieve the private training goal.

[0088] S2: analyzing the training demand to obtain a training demand structure and a training demand mode.

[0089] Specifically, the training demand structure includes the organization and arrangement of each component in the training process. That is, the training training service provided by the generative large model needs to meet the training demand structure in the process.

[0090] Specifically, the training demand mode includes a specific method or strategy adopted in the training process. That is, a specific method or strategy needs to be adopted in the process of providing the training training service by the generative large model.

[0091] S3: generating the intelligent training training service based on the training demand structure and the training demand mode by the generative large model.

[0092] In some embodiments, the scenario feature of the user's training demand is a private training scenario; the private training scenario includes a private role, a coach role, and a private training goal. Then, generating the intelligent training training service based on the training demand structure by the generative large model includes: the generative large model provides the user in the private role with the intelligent training training service in the coach role to achieve the private training goal.

[0093] As shown in the figure, the generative large model provides the user in the private role with the intelligent training training service in the coach role includes steps S31-34. Figure 4

[0094] S31: the generative large model provides the user with the corresponding standard language structure and answer demonstration based on the training theme.

[0095] Specifically, each training theme has a corresponding standard language structure and an answer demonstration meeting the standard language structure. Then, when a specific training theme is determined, the generative large model provides the user with the corresponding standard language structure and answer demonstration. The "standard language structure and answer demonstration" in this application is not only the basis for the user's practice and learning, but also the basis for the automatic checking, calibration and guidance of the user by the application (i.e. intelligent coach).

[0096] ​For example, the standard language structure for customer service to guide customers in a business scenario is: direction + distance + landmark.

[0097] For example, the standard language structure for resolving customer dissatisfaction in a business scenario is: apology + fact + suggestion + action support.

[0098] For example, the basic structure of customer service communication in a business scenario is: transaction + emotion.

[0099] For example, the basic structure of experience marketing is: invitation + guidance + resonance.

[0100] The answer to the standard language structure for customer service to guide customers in a business scenario is: straight ahead / 40 meters / right side / red billboard / below.

[0101] In some embodiments, the standard language structure can also include keywords.

[0102] For example, the language to invite customers to participate in the experience: You can, you can, you can try it…

[0103] For example, the guiding language to guide customers to experience: touch, listen, look, knock, smell, pull…

[0104] For example, to trigger user experience resonance: Have you? Isn't it?

[0105] S32: The generative large model retrieves an infinite number of scene questions based on the training theme and preset categories and conditions.

[0106] Specifically, in the private coaching scenario, for each training theme under each private coaching goal, the categories and conditions are preset in advance, such as the scene of setting the question, the required elements (scene question elements), the limit of the number of words, and the content that cannot appear. The generative large model retrieves an infinite number of scene questions corresponding to the training theme according to the preset categories and conditions, according to the setting question demonstration input in advance, thereby providing users with diversified simulation training scenarios and effectively achieving the task goal.

[0107] In some embodiments, the difficulty of the scene question under the same training theme can also be set. As the number of scene questions answered by the user increases, the difficulty of the scene questions output by the generative large model will gradually increase to increase the challenge for the user.

[0108] Specifically, the elements possessed by the scene question include any one or several of the task, time, place, character, character relationship, character emotional state, and key event.

[0109] It should be noted that the generative large model starts to capture endless scene topics based on the starting instruction of the user. When the user does not issue a starting instruction or issues irrelevant instruction language, the generative large model does not respond.

[0110] S33: The generative large model outputs the scene topic based on the preset normative instruction structure.

[0111] Specifically, the normative instruction structure includes polite expression requirements, encouraging expression requirements, introductory guiding expressions at the beginning, and open guiding expressions at the end. The generative large model outputs the scene topic based on the normative instruction structure preset in advance.

[0112] S34: The generative large model identifies the user's response to the scene topic, matches the standard language structure, and outputs a response evaluation.

[0113] Specifically, after the user responds to the scene topic, the generative large model automatically extracts the language structure and keywords in the response content and matches them with the standard language structure to output a response evaluation. That is, the generative large model determines whether the user's response conforms to the standard language structure and outputs a response evaluation accordingly.

[0114] When the generative large model determines that the response lacks the standard language structure, the generative large model triggers a teaching mode and outputs a prompt to guide the user to continue the complete response content until the response meets the standard language structure.

[0115] When the generative large model determines that the response meets the standard language structure, the generative large model triggers a question mode, outputs the next scene topic, and can reply to the user's praise evaluation, encouraging the user to continue.

[0116] In the private coaching scenario, the training demand structure includes the standard language structure and answer demonstration corresponding to the training theme, and the corresponding evaluation response content can be set as needed, and the present application is not limited thereto.

[0117] Thus, in the private coaching scenario, the generative large model provided by the present application serves as a coach and can provide one-on-one targeted training for the user. According to the private coaching goal selected by the user, the generative large model generates a training theme and provides endless scene topics to achieve the private coaching goal. Moreover, the generative large model also identifies and analyzes the user's answer content and provides timely training feedback and correction suggestions for the user.

[0118] Please refer to Figures 5a to 5c as shown, Figures 5a to 5cThe interface display figure of the intelligent training assistance provided by the generative large model in the private training scene is shown. The user can select the corresponding training theme, and the generative large model provides the corresponding standard language structure and answer demonstration. At this time, the private training interface is entered, the user answers the scene question, and the generative large model responds and evaluates.

[0119] In some embodiments, the scene feature of the training needs of the user is an assistant trainer scene; the assistant trainer scene includes a trainer role, a trainer assistant role, and a training target. Then, the intelligent training service is generated by the generative large model based on the training needs structure and the training needs mode, which includes that the generative large model provides intelligent training assistance services to the user in the trainer assistant role to the trainer role to achieve the training target.

[0120] As shown in Figure 6 The generative large model provides intelligent training assistance services to the user in the trainer assistant role to the trainer role includes steps S35-S37.

[0121] S35: The generative large model extracts training key information based on the training needs.

[0122] In some embodiments, the training key information includes a training course name, a trained industry, a trained enterprise, a student post, a number of students, a training duration, and a focus solving problem.

[0123] S36: The generative large model generates first training materials based on the training needs structure and the training key information.

[0124] In some embodiments, the first training materials need to have specific components, including the organization and arrangement of components, that is, the training needs structure. At this time, after obtaining the training key information, the first training materials are generated based on the training needs structure and the training key information, and the first training materials are rough information frameworks.

[0125] S37: The generative large model extracts endless relevant information based on the training needs mode and the preset category and condition, and supplements the first training materials based on the relevant information to generate second training materials that conform to the preset content structure.

[0126] In some embodiments, the generative large model adopts a specific method or strategy to extract endless relevant information under the preset category and condition, supplement the first training materials, and finally generate second training materials that conform to the preset content structure, and the second training materials are complete training content.

[0127] Further, the generative large model provides intelligent training assistance services to the user in the role of the trainer in the role of the trainer assistant, including (A) (B) (C) (D) (E) embodiments. Specifically, it includes:

[0128] (A) The generative large model provides a course outline generation service to the user in the role of the trainer in the role of the trainer assistant, including:

[0129] The generative large model extracts course key information based on the training needs; for example, the course key information includes training course name, training industry, training enterprise, student position, student number, training duration, focus problem solving;

[0130] The generative large model generates course information based on the training needs structure and the course key information; for example: the course information includes: course name, course problem solving, course goal, student object, course duration, training method;

[0131] The generative large model extracts an infinite amount of course-related information based on the training needs pattern and the preset course category and conditions, and supplements the course information based on the course-related information to generate course content; for example, the course content includes the title of each chapter, knowledge point content, training method, core point or / and tool method;

[0132] The generative large model generates a training course outline that meets the preset content structure based on the course information and the course content.

[0133] (B) The generative large model provides a courseware generation service to the user in the role of the trainer in the role of the trainer assistant, including:

[0134] The generative large model extracts knowledge point key information based on the training needs; the training needs include a preset course outline or a training course outline generated by the generative large model;

[0135] The generative large model generates knowledge point information based on the training needs structure, the training course outline, and the knowledge point key information;

[0136] The generative large model extracts an infinite amount of knowledge point-related information based on the training needs pattern and the preset knowledge point category and conditions, and supplements the knowledge point information based on the knowledge point-related information to generate knowledge point content;

[0137] The generative large model generates course knowledge points that meet the preset content structure based on the knowledge point information and the knowledge point content.

[0138] Further, when the user selects the training theme as "generating courseware", the generative large model provides a service of generating courseware, at this time, the generative large model generates courseware according to the user's pre-input course outline or the course outline generated by the generative large model in addition to generating courseware according to the key information of knowledge points (training course name, trained industry, trained enterprise, student position, student number, training duration, focus problem solving). That is, when the generative large model only obtains the key information of knowledge points (training course name, trained industry, trained enterprise, student position, student number, training duration, focus problem solving), in order to realize the service of generating courseware, the generative large model will first generate a course outline, and then further generate courseware based on it.

[0139] (C) The generative large model provides a service of generating content materials to the user in the role of a trainer in the role of a trainer assistant, including:

[0140] The generative large model extracts material key information based on the training demand; the training demand includes a preset course outline or the training course outline generated by the generative large model;

[0141] The generative large model generates material information based on the training demand structure, the training course outline and the material key information; the material information includes course knowledge points, case stories and tool methods;

[0142] The generative large model extracts infinite material related information based on the training demand mode and the preset material category and condition, and supplements the material information based on the material related information to generate material content;

[0143] The generative large model generates training materials conforming to the preset content structure based on the material information and the material content.

[0144] Further, the material content includes knowledge point content, training interaction and the like. That is, when the generative large model only obtains the key information of knowledge points (training course name, trained industry, trained enterprise, student position, student number, training duration, focus problem solving), the generative large model can first generate training interaction, knowledge point content and the like, and then further generate material content.

[0145] (D) The generative large model provides a service of generating course test questions to the user in the role of a trainer in the role of a trainer assistant, including:

[0146] The generative large model extracts test question key information based on the training demand; the training demand also includes user-entered knowledge points or the course knowledge points generated by the generative large model;

[0147] The generative large model generates test question information based on the training demand structure, the test question key information, and the course knowledge points in the material information;

[0148] The generative large model extracts infinite test question related information based on the training demand mode and the preset test question category and condition, supplements the test question information based on the test question related information, and generates test question content;

[0149] The generative large model generates training test questions that cover the course knowledge points and meet the preset content structure based on the test question information and the test question content.

[0150] (E) The generative large model provides the training interaction service to the user in the role of a trainer assistant, including:

[0151] The generative large model extracts interactive key information based on the training demand; the training demand includes a preset course outline or the training course outline generated by the generative large model;

[0152] The generative large model generates interactive information based on the training demand structure, the training course outline, and the interactive key information; the interactive information includes opening interaction, process interaction, and closing interaction;

[0153] The generative large model extracts infinite interactive related information based on the training demand mode and the preset interactive category and condition, supplements the interactive information based on the interactive related information, and generates interactive content;

[0154] The generative large model generates training interaction that meets the preset content structure based on the interactive information and the interactive content.

[0155] Taking the "(A) course outline generation service" as an example, the course outline key information includes the training course name, the trained industry, the trained enterprise, the student position, the number of students, the training duration, and the focus problem to be solved. The training demand structure can include "the problem to be solved by the course", and according to the training demand structure "the problem to be solved by the course" and the course outline key information "the focus problem to be solved", the course outline information "the problem to be solved by the course" can be generated. For "the problem to be solved by the course", the course outline related information needs to be extracted under the category and condition of "giving the value point of the course, the economic value brought to the enterprise, and the social reputation value" strategy, and the course outline information "the problem to be solved by the course" is supplemented to generate the corresponding course outline content. Taking this as an example, all course outline information can be supplemented to generate course outline content, and finally generate a training course outline that meets the preset content structure.

[0156] In some embodiments, the generation of the training course outline conforming to the preset content structure is displayed as:

[0157] "From the content input by the user, starting from the standard prompt, the prompt demonstrates as follows: For you, the xxx company xxx post xxx hour of the "xxxx" training course outline is as follows:

[0158] First chapter

[0159] Section 1

[0160] Case:

[0161] Tools:

[0162] Viewpoint:"

[0163] It should be noted that in the above embodiments, the generative large model is actually not completely independent when providing services such as generating course outlines, training materials, training interactions, courseware, and course test questions. In other words, the generative large model can provide a certain generation service on the basis of other generation services. For example, the generative large model can further generate courseware on the basis of generating course outlines.

[0164] It should be noted that the above content is only used as an example for illustration. In the scenario of assisting the trainer, the training demand structure, the training demand mode, and the course outline category and condition can be set according to the needs, and the present application is not limited thereto.

[0165] Please refer to Figures 7a to 7b As shown, Figures 7a to 7b The interface display diagram of the intelligent training assistance service provided by the generative large model in the scenario of assisting the trainer is shown. The user can select the corresponding training theme and input the training key information, and the generative large model generates the second training material conforming to the content structure according to the training demand structure, the training demand mode, and the preset category and condition.

[0166] Therefore, the present application can "one-click generate" accurate course outlines, training materials, training interactions, courseware, and course test questions for trainers based on the key information input by the user and the selected training theme, thereby greatly improving the desk work efficiency of the trainer, reducing the training preparation work burden, saving more rest time, thinking and creating time, and more fully investing in the training work. For the purchasing enterprise, it can also obtain accurate and efficient business training.

[0167] In some embodiments, the intelligent training method based on the generative large model provided by the present application further includes: obtaining training data of a user, and providing the training data to the user based on a trigger instruction of the user.

[0168] Specifically, the training data includes any one or more of training duration, practice accuracy, training ranking, and training time period. Please refer to Figures 8a to 8c as shown, Figures 8a to 8c a training data management interface schematic diagram is shown. For training data, it can be displayed through a tree chart, a line chart, etc., and the present application is not limited thereto.

[0169] Thus, the present application can enable users to obtain more real, direct and accurate training result data, so that enterprises can make further training decisions based on the training data.

[0170] As Figure 9 shown, the present application also provides a model training method. Please refer to Figure 9 shown, the training method includes steps S51-S56.

[0171] S51: Obtain a training theme; the training theme corresponds to a model role feature and a task target.

[0172] Specifically, the training theme is extracted based on experience or real enterprise situation, and the training theme corresponds to the model role feature and the task target.

[0173] In some embodiments, the model role feature is a question setter and an evaluator, and the task target includes experience marketing, resistance resolution, guided parking, persuasion behavior, guide reception, complaint resolution, new product mastery, offline meeting invitation, expression of praise, and expression of refusal. At this time, the training theme corresponds to the task target, which is a key touch point task corresponding to the task target. For example, when the task target is experience marketing, the training theme is a key touch point task of the task target of experience sales, such as inviting customers to participate in experience, inviting customers to experience, and arousing user experience empathy.

[0174] In other embodiments, the model role feature is a training assistant, and the task target includes conducting enterprise business training. At this time, the training theme is generating a course outline, generating a training courseware display, generating training content materials, generating training activities, and generating training test questions corresponding to the task target.

[0175] S52: Obtain training elements corresponding to the training theme; the training elements include a category and a condition of a scene question corresponding to the training theme, a standard language structure, or a content structure corresponding to the training theme.

[0176] Specifically, when the model role feature is set to a question setter and an evaluator, the category and condition of a scene question corresponding to the training theme, and the standard language structure for evaluation are set.

[0177] For example, the standard language structure for guiding customers in a business scenario is set as: direction + distance + landmark.

[0178] For example, the standard language structure for resolving customer dissatisfaction in a business scenario is set as: apology + fact + suggestion + action support.

[0179] For example, the basic structure for customer service communication in a business scenario is set as: transaction + emotion.

[0180] For example, the basic structure for experience marketing is set as: invitation + guidance + resonance.

[0181] For example, the category and conditions of the scenario topic are set as the scenario of the question, the required elements, the word limit, and the content that cannot appear.

[0182] Specifically, when the model character feature is set as a training assistant, the content structure corresponding to the training topic is set.

[0183] For example, taking the training topic "generate course outline" as an example, the content structure is:

[0184] "Start with the standard prompt in the user's input content, and the prompt demonstration is as follows: For you to output xxx company xxx position xxx hours of 《xxxx》 training course outline as follows:

[0185] First chapter

[0186] Section 1

[0187] Case:

[0188] Tools:

[0189] View:"

[0190] For example, taking the training topic "generate courseware" as an example, the content structure is:

[0191] Start with the standard prompt in the user's input content, and the prompt demonstration is as follows: For you to output xxx company xxx position xxx hours of 《xxxx》 training course knowledge points as follows, including knowledge point introduction, specific operation and usage instructions.

[0192] Format demonstration:

[0193] First part: Knowledge point introduction

[0194] Second part: Tool and method introduction

[0195] Third part: Tool and method operation instruction

[0196] The last sentence of the standard prompt is: Do you need any supplements? "

[0197] It should be noted that the content structure can be set as needed, and the present application is not limited thereto.

[0198] S53: Set the normative instruction structure of the user training model.

[0199] Specifically, the normative instruction structure includes politeness requirement, encouragement requirement, introductory phrase at the beginning, and open-ended introductory phrase at the end.

[0200] For example, the introductory phrase at the beginning is set as "Please start answering the questions".

[0201] S54: Construct a model training set based on the training theme, the training elements and the normative instruction structure.

[0202] S55: Train the generative large model based on the model training set to obtain a trained generative large model.

[0203] S56: Obtain a training demonstration set, which includes a positive demonstration set, a missing demonstration set, a correction and supplement demonstration set, and a guidance demonstration set.

[0204] S57: Based on the training demonstration set, the trained generative large model is respectively trained in a positive, negative, correction and supplement, and guidance demonstration manner to optimize the cognition of the generative large model, and generate an optimized generative large model, i.e. an intelligent training model.

[0205] Specifically, the training demonstration set further provides positive demonstration, missing demonstration, correction and supplement demonstration, and guidance demonstration to help the generative large model establish cognition and generate an optimized generative large model.

[0206] When training the generative large model with prompt words, different parameters can be adjusted to meet the training needs of different application scenarios.

[0207] When it is a private coaching scenario, the parameters of the generative large model can be adjusted.

[0208] When it is an assistant trainer scenario, the parameters of the generative large model can be adjusted.

[0209] In an embodiment of the present application, the present application uses the tokens_to_generate parameter of the generative text model to specify how many tokens the model should generate after being given input. Tokens can be words, characters, subwords or other language units, depending on the tokenization method of the model. Specifically, it includes:

[0210] Determining the generation length: First, the desired length of the generated text needs to be determined. This can be based on specific application requirements, such as generating a short sentence, a paragraph of description, or a complete story.

[0211] Setting the parameters: When calling the API or function of the model, set tokens_to_generate to the desired number of tokens to generate.

[0212] Adjusting the quality of generation: Sometimes, increasing or decreasing the value of tokens_to_generate can affect the quality and relevance of the generated text. By providing more tokens, the application can provide more detailed output, but it may also increase the risk of generating irrelevant or logically inconsistent text.

[0213] Considering the context limit: When setting tokens_to_generate, the application considers the context window size of the model. If the number of generated tokens approaches or exceeds the maximum context length of the model, it may cause the generated text to be disconnected from the previous text.

[0214] Testing and optimization: In practical applications, the value of tokens_to_generate is adjusted multiple times to find the optimal generation length, which depends on specific requirements and the performance of the model. By reasonably setting tokens_to_generate, the application can more effectively control the output of the generative text model, making it more suitable for the application scenario and user requirements of the application.

[0215] In an embodiment of the application, the application uses the top_p parameter in the generative text model to control the diversity and randomness of the generation process, aiming to generate more diverse and interesting text while avoiding generating overly common or highly predictable text. For example, when top_p = 0.9, it means that the model will consider the top 90% of words with the highest probability when generating each word. By adjusting the value of top_p, the application observes and tests the changes in generated text to find the most suitable setting for the requirements and adjust the generation effect.

[0216] The application uses the top_p parameter together with the temperature parameter to further control the generation process. The temperature parameter controls the randomness in the generation process, with lower temperature values making the generated text more inclined to high-probability words, and higher temperature values increasing randomness.

[0217] Through experiments and optimization, the application finds the most suitable setting for specific requirements. Different text generation tasks may require different top_p settings.

[0218] The present application considers the importance of context when generating text. In situations that require high relevance and logical coherence, more careful settings of top_p are needed. By using the top_p parameter reasonably, the present application can achieve better control in generative text models, generating interesting and expected text.

[0219] In an embodiment of the present application, the present application uses the prompt parameter in the generative text model to determine the text input into the model while guiding the model to generate specific output. The prompt can be a question, a description, the beginning of a story, or any type of text that aims to stimulate the model to produce relevant and coherent responses.

[0220] The role of the prompt described in the present application includes:

[0221] Guiding the generation of content: The prompt provides the context and direction for the model to generate text. It helps the model understand what kind of text needs to be generated.

[0222] Control the theme of the output: Through a well-designed prompt, the model can be guided to generate text of a specific theme or style.

[0223] Improve relevance: A suitable prompt can improve the relevance and accuracy of the generated text, making it more in line with the user's expectations.

[0224] The process of designing an effective prompt in the present application includes:

[0225] Clearly define the goal: Before designing the prompt, clearly define what kind of text you want the model to generate. This will help design a more targeted prompt.

[0226] Provide sufficient context: Ensure that the prompt contains enough information so that the model can understand the background knowledge needed to generate the text.

[0227] Use clear and specific language: Avoid ambiguous and ambiguous expressions, which can help the model more accurately understand the intent.

[0228] Consider the capabilities and limitations of the model: Different models may react differently to different types of prompts. Understand the characteristics of the model and design a prompt suitable for that model.

[0229] Test and iterate: Try different prompts, observe the model's response, and then adjust the prompt as needed to optimize the results.

[0230] For example, suppose you are using a text generation model to generate a sequel to a story, you can design a prompt as follows:

[0231] Simple prompt: "Once upon a time, in a small village..."

[0232] Detailed prompt: "In a small, peaceful village surrounded by dense forests, a young girl named Ella discovered a hidden path that led to a mysterious cave. As she ventured deeper into the cave, she found..."

[0233] The combination of prompts with model parameters will be used in actual applications, combining prompts with other parameters of the model (such as top_p, temperature, max_tokens, etc.) to achieve more fine-grained control and optimize the generated results.

[0234] By carefully designing and adjusting the prompts, you can effectively use the generative text model to meet specific needs, whether it's content creation, data augmentation, or other applications. The protection scope of the intelligent training method based on the generative large model described in the embodiments of the present application is not limited to the execution order of the steps listed in the embodiments. Any scheme achieved by adding, replacing or replacing steps of the prior art according to the principles of the present application is included in the protection scope of the present application.

[0235] The embodiments of the present application also provide an intelligent training system based on a generative large model, which can implement the intelligent training method based on the generative large model described in the present application. However, the implementation device of the intelligent training system based on the generative large model described in the present application includes but is not limited to the structure of the intelligent training system based on the generative large model listed in the embodiments. Any structural deformation and replacement of the prior art according to the principles of the present application is included in the protection scope of the present application.

[0236] As shown in Figure 10 In an embodiment, the intelligent training system based on the generative large model of the present application includes an acquisition module 41, an analysis module 42 and a training module 43.

[0237] The acquisition module 41 is used to acquire the training needs of the user.

[0238] The analysis module 42 is configured to analyze the training demand, obtain a training demand structure and a training demand mode; the training demand structure includes an organization and arrangement of each component in the training process; and the training demand mode includes a specific method or strategy used in the training process.

[0239] The training module 43 is configured to generate an intelligent training service based on the training demand structure and the training demand mode by using the generative large model.

[0240] The structure and principle of the obtaining module 41, the analysis module 42 and the training module 43 correspond to the steps of the intelligent training method based on the generative large model, and thus will not be described here.

[0241] In the several embodiments provided in the present application, it should be understood that the disclosed system, device or method can be implemented in other ways. For example, the device embodiments described above are only schematic. The division of the modules / units is only a logical function division. There can be another division manner in actual implementation. For example, a plurality of modules or units 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 shown or discussed modules can be indirect coupling or communication connection through some interfaces, devices or modules, and can be electrical, mechanical or other forms.

[0242] The modules / units described as separate components can or can not be physically separate, and the components shown as modules / units can or can not be physical modules, i.e., can be located in one place or distributed on a plurality of network units. Some or all of the modules / units can be selected according to actual needs to achieve the purpose of the embodiments of the present application. For example, the functional modules / units in each embodiment of the present application can be integrated into one processing module, or each module / unit can be physically separate, or two or more modules / units can be integrated into one module / unit.

[0243] Those of ordinary skill in the art should further appreciate that the units and algorithm steps of the examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, computer software or a combination of both. In order to clearly illustrate the interchangeability of hardware and software, the components and steps of the examples have been described in general terms in the above description. Whether the functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. A skilled person can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0244] The embodiments of the present application further provide a computer readable storage medium. Those skilled in the art can understand that all or part of the steps of the method described above can be instructed by a program to complete the processor, and the program can be stored in a computer readable storage medium. The storage medium is a non-transitory medium, such as a random access memory, a read-only memory, a flash memory, a hard disk, a solid state disk, a magnetic tape, a floppy disk, an optical disc and any combination thereof. The storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server, a data center and the like, which includes one or more available media sets. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a magnetic tape), an optical medium (for example, a digital video disc (DVD)) or a semiconductor medium (for example, a solid state disk (SSD)) and the like.

[0245] The embodiments of the present application further provide an electronic device. The electronic device comprises a processor and a memory.

[0246] The memory is used to store a computer program.

[0247] The memory comprises a ROM, a RAM, a disk, a U disk, a memory card or an optical disc and the like various media that can store program codes.

[0248] The processor is connected with the memory, and is used to execute the computer program stored in the memory, so that the electronic device executes the intelligent training method based on the generative large model described above.

[0249] Preferably, the processor can be a general processor, including a central processing unit (CPU), a network processor (NP) and the like; and can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components.

[0250] As Figure 11As shown, the electronic device of the present application is in the form of a general- purpose computing device. The components of the electronic device can include, but are not limited to, one or more processors or processing units 51, a memory 52, and a bus 53 that couples various system components including the memory 52 to the processing unit(s) 51.

[0251] The bus 53 represents one or more of any of several types of bus structures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration bus (e.g., AGP bus) and a processor or local bus using any of a variety of bus architectures. By way of example, these architectures include Industry Standard Architecture (ISA) bus, Micro Channel Architecture (MCA) bus, Enhanced ISA bus, Video Electronics Standards Association (VESA) local bus, and Peripheral Component Interconnect (PCI) bus.

[0252] The electronic device typically includes a variety of computer system readable media. Such media can be any available media that is located either internally or externally to the electronic device. It includes storage of removable and non-removable, volatile and non-volatile computer system storage media.

[0253] The memory 52 can include computer system readable media in the form of volatile memory, such as random access memory (RAM) 521 and / or cache memory 522. The electronic device can further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, a storage system 523 can be provided for reading from and writing to non-removable, non-volatile magnetic media (e.g., a "hard drive"). Figure 11 Not shown, a removable / non-removable, volatile / non-volatile computer system storage medium can be a magnetic disk drive that reads from and / or writes to a non-removable, non-volatile magnetic media (not shown and typically called a "hard drive"). Figure 11 In these instances, each drive can be connected to the bus 53 by one or more data media interfaces. The memory 52 can include at least one program product having a set (e.g., at least one) of program modules that are configured to carry out the functions of embodiments of the present application.

[0254] A program / utility 524, having a set (at least one) of program modules 5241, can be stored in memory 52 by way of example, such program modules 5241 include an operating system, one or more application programs, other program modules, and program data, each of which or a combination can include implementation of a networking environment. Program modules 5241 generally carry out the functions and / or methodologies of embodiments of the present application as described herein.

[0255] The electronic device can also communicate with one or more external devices such as a keyboard or a pointing device, displays, etc.; other devices that enable a user to interact with the electronic device; and / or any devices (e.g., network card, modem, etc.) that enable the electronic device to communicate with one or more other computing devices. Such communication can occur via Input / Output (I / O) interface 54. Still yet, the electronic device can communicate with one or more networks, such as a local area network (LAN), a wide area network (WAN), and / or the Internet, through network adapter 55. As depicted, network adapter 55 communicates with the other components of the electronic device via bus 53. It should be appreciated that although not shown, other hardware and / or software components could be used in conjunction with the electronic device. These components, as well as the electronic device, generally communicate through the use of appropriate Figure 11 application-specific component(s) 56. Additionally, it should be appreciated that the electronic device has the ability to operate in a networked environment using logical and / or physical connections made through a wired or wireless communication subsystem to one or more other computers, such as computing devices 80 across a network 81 as illustrated in Figure 8.

[0256] Embodiments of the present application can also provide a computer program product including one or more computer instructions. When the computer instructions are loaded and executed on a computing device, all or part of the processes or functions described in embodiments of the present application are generated. The computer instructions can be stored in a computer readable storage medium, or transferred from one computer readable storage medium to another computer readable storage medium, for example, the computer instructions can be transferred from one website site, computer or data center to another website site, computer or data center through wired (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.) mode.

[0257] The computer program product is executed by a computer, and the computer executes the method described in the foregoing method embodiments. The computer program product can be a software installation package, and in the case of needing to use the foregoing method, the computer program product can be downloaded and executed on the computer.

[0258] The application provides an intelligent training method, system, medium, device and model training method based on a generative large model, which can provide one-to-one training services for users, effectively provide targeted training services around the task target, and users can receive instant feedback and correction. At the same time, it meets the user's privacy needs and can also meet the actual application scenarios. Secondly, the enterprise training model can provide automatic generation of training content that meets different training themes for users, reducing the time and labor cost paid by users for business training, and providing more standardized training services. For enterprises, the probability of obtaining accurate and efficient training is greatly improved. Through the application, individuals can effectively improve their business skills, improve their work efficiency, quickly pass the employee training period, and maximize their work value; enterprises can fully focus on business goals to provide training, solve the actual problem of improving the competence of external interaction of the post, greatly shorten the competence growth cycle of the post ability of the employees, reduce the cost of human, material and financial resources in the process of personnel training, and the loss of training results caused by the loss of enterprise employees, and realize the cost reduction and benefit increase. In addition, the application provides a training data recording service, which can provide users with more real, direct and accurate training result data, so that enterprises can make further training decisions based on the training data.

[0259] The description of the flow or structure corresponding to each of the above figures has its own emphasis, and the parts not described in detail in a certain flow or structure can be referred to the related description of other flows or structures.

[0260] The above embodiments only exemplarily illustrate the principles and effects of the application, and are not used to limit the application. Any person skilled in the art can modify or change the above embodiments without departing from the spirit and scope of the application. Therefore, all equivalent modifications or changes completed by those skilled in the art without departing from the spirit and technical idea of the application should be covered by the claims of the application.

Claims

1. An intelligent training training method based on a generative large model, characterized in that, The application relates to a method for generating intelligent training services based on a user's training demand. The method comprises the following steps: acquiring the user's training demand; analyzing the training demand to obtain a training demand structure and a training demand mode; the training demand structure comprises the organization and arrangement of various components in the training process; the training demand mode comprises specific methods or strategies adopted in the training process; 2.The intelligent training training method based on a generative large model according to claim 1, wherein, generating intelligent training services based on the training demand structure and the training demand mode by using a generative large model. The method for acquiring the user's training demand comprises the following steps: acquiring scene characteristics of the training demand; the scene characteristics comprise a first role characteristic of the user, a second role characteristic of a role interacting with the user and a task target; 3.The intelligent training training method based on a generative large model according to claim 2, characterized in that, generating a corresponding matched training theme based on the scene characteristics of different contents; that is, generating a training theme based on the first role characteristic of the user, the second role characteristic of the role interacting with the user and the task target. The scene characteristics of the user's training demand are private training scene characteristics; the private training scene characteristics comprise a private role, a coach role and a private training target. The method for generating intelligent training services based on the training demand structure by using the generative large model comprises the following steps: the generative large model provides intelligent training services for the user in the coach role to the private role to achieve the private training target; specifically, the method comprises the following steps: the generative large model provides corresponding standard language structures and answer demonstrations for the user based on the training theme; the generative large model extracts endless scene questions based on the training theme and preset categories and conditions; the scene questions comprise any one or several of a task, a time, a place, a character, a character relationship, a character emotional state and a key event; the generative large model outputs the scene questions based on a preset normative instruction structure; the generative large model identifies the user's response to the scene questions, matches the standard language structure and outputs a response evaluation; wherein, 4.The intelligent training training method based on a generative large model according to claim 2, characterized in that, when the response does not conform to the standard language structure and there is a structure defect, the generative large model triggers a teaching mode, outputs a prompt for guiding the user to continue the complete response content until the response conforms to the standard language structure; when the response conforms to the standard language structure, the generative large model triggers a question mode and outputs the next scene question. The scene characteristics of the user's training demand are assistant trainer scene characteristics; the assistant trainer scene characteristics comprise a trainer role, an assistant trainer role and a training target. The method for generating intelligent training services based on the training demand structure and the training demand mode by using the generative large model comprises the following steps: the generative large model provides intelligent training services for the user in the assistant trainer role to the trainer role to achieve the training target; specifically, the method comprises the following steps: the generative large model extracts training key information based on the training demand; the generative large model generates first training materials based on the training demand structure and the training key information; the generative large model extracts endless related information based on the training demand mode and preset categories and conditions, supplements the first training materials based on the related information and generates second training materials conforming to a preset content structure. 5.The intelligent training training method based on a generative large model according to claim 4, characterized in that, The generative large model provides intelligent training assistance services to the user in the role of the trainer in the role of the trainer assistant also includes: the generative large model provides course outline generation services to the user in the role of the trainer in the role of the trainer assistant, including: The generative large model extracts key information based on the training needs; The generative large model generates course information based on the training needs structure and the key information of the course; The generative large model extracts endless course-related information based on the training needs mode and the preset course category and condition, and supplements the course information based on the course-related information to generate course content; The generative large model generates a training course outline that meets the preset content structure based on the course information and the course content. 6.The intelligent training training method based on a generative large model according to claim 5, characterized in that, The generative large model provides intelligent training assistance services to the user in the role of the trainer in the role of the trainer assistant also includes: the generative large model provides courseware generation services to the user in the role of the trainer in the role of the trainer assistant, including: The generative large model extracts knowledge point key information based on the training needs; the training needs include a preset course outline or a training course outline generated by the generative large model; The generative large model generates knowledge point information based on the training needs structure, the training course outline and the knowledge point key information; The generative large model extracts endless knowledge point-related information based on the training needs mode and the preset knowledge point category and condition, and supplements the knowledge point information based on the knowledge point-related information to generate knowledge point content; The generative large model generates course knowledge points that meet the preset content structure based on the knowledge point information and the knowledge point content. 7.The intelligent training training method based on a generative large model according to claim 5, wherein, The generative large model provides intelligent training assistance services to the user in the role of the trainer in the role of the trainer assistant also includes: the generative large model provides content material generation services to the user in the role of the trainer in the role of the trainer assistant, including: The generative large model extracts material key information based on the training needs; the training needs include a preset course outline or a training course outline generated by the generative large model; The generative large model generates material information based on the training needs structure, the training course outline and the material key information; the material information includes course knowledge points, case stories and tool methods; The generative large model extracts endless material-related information based on the training needs mode and the preset material category and condition, and supplements the material information based on the material-related information to generate material content; The generative large model generates training materials that meet the preset content structure based on the material information and the material content. 8.The intelligent training training method based on a generative large model according to claim 6, characterized in that, The generative large model provides intelligent training assistance services to the user in the role of the trainer in the role of the trainer assistant also includes: the generative large model provides course test question generation services to the user in the role of the trainer in the role of the trainer assistant, including: The generative large model extracts test question key information based on the training requirements; the training requirements also include user-entered knowledge points or the course knowledge points generated by the generative large model; The generative large model generates test question information based on the training requirement structure, the test question key information, and the course knowledge points in the material information; The generative large model extracts infinite test question related information based on the training requirement mode and the preset test question category and conditions, and supplements the test question information based on the test question related information to generate test question content; The generative large model generates training test questions that cover the course knowledge points and meet the preset content structure based on the test question information and the test question content. 9.The intelligent training training method based on a generative large model according to claim 4, characterized in that, The generative large model provides intelligent training assistance services to users in the role of a trainer assistant also includes: the generative large model provides training interaction generation services to users in the role of a trainer assistant, including: The generative large model extracts interaction key information based on the training requirements; the training requirements include a preset course outline or the training course outline generated by the generative large model; The generative large model generates interaction information based on the training requirement structure, the training course outline, and the interaction key information; the interaction information includes opening interactions, process interactions, and closing interactions; The generative large model extracts infinite interaction related information based on the training requirement mode and the preset interaction category and conditions, and supplements the interaction information based on the interaction related information to generate interaction content; The generative large model generates training interactions that meet the preset content structure based on the interaction information and the interaction content.

10. An intelligent training training system based on a generative large model, characterized by, The system is configured to: An acquisition module for acquiring a user's training requirements; An analysis module for analyzing the training requirements to obtain a training requirement structure and a training requirement mode; the training requirement structure includes the organization and arrangement of each component in the training process; the training requirement mode includes specific methods or strategies used in the training process; A training module for generating intelligent training services using a generative large model based on the training requirement structure and the training requirement mode.

Images