Cboth case generation method and device, electronic equipment and storage medium

Abstract

The embodiment of the invention discloses a copywriting generation method and device, electronic equipment and a storage medium, relates to the technical field of computers, and can effectively improve the quality of generated copywriting. The method comprises the following steps: acquiring copywriting demand description information; constructing a copywriting generation plan according to the copywriting demand description information, wherein the copywriting generation plan is used for generating a target copywriting; the copywriting generation plan comprises a plurality of subtasks executed in sequence, and each subtask is used for processing one copywriting attribute of the target copywriting through a corresponding copywriting processing model; and generating the target copywriting according to the copywriting generation plan. The method and the device can be suitable for copywriting generation scene based on artificial intelligence.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a document generation method, device, electronic device and storage medium. Background Art

[0002] At present, AI-based copywriting generation tools can generate various copywriting for users based on the prompts entered by users, using a large language model as the core generation engine. However, for some more complex copywriting requirements, many of the generated copywriting cannot satisfy users and the copywriting quality is poor. Summary of the invention

[0003] In view of this, the embodiments of the present application provide a copy generation method, device, electronic device and storage medium, which can effectively improve the quality of generated copy.

[0004] In a first aspect, an embodiment of the present application provides a copywriting generation method, comprising: obtaining copywriting requirement description information; constructing a copywriting generation plan according to the copywriting requirement description information, wherein the copywriting generation plan is used to generate a target copywriting; the copywriting generation plan comprises a plurality of subtasks executed sequentially, wherein each of the subtasks is used to process a copywriting attribute of the target copywriting through a corresponding copywriting processing model; and generating the target copywriting according to the copywriting generation plan.

[0005] In one embodiment, constructing a copy generation plan based on the copy requirement description information includes: determining the copy type, user requirements, and target audience of the target copy based on the copy requirement description information; and constructing the copy generation plan based on the copy type, user requirements, and target audience of the target copy.

[0006] In one embodiment, constructing the copy generation plan according to the copy type, user needs, and target audience of the target copy includes: constructing prompt words of the copy generation plan according to the copy type, user needs, and target audience of the target copy; providing the prompt words of the copy generation plan to a copy generation plan construction model, and obtaining the copy generation plan output by the copy generation plan construction model based on the prompt words.

[0007] In one embodiment, after obtaining the copy generation plan output by the copy generation plan construction model based on the prompt words, the method further includes: providing the copy generation plan to the user, receiving a modification instruction from the user for the copy generation plan, and adjusting the copy generation plan according to the modification instruction to obtain an adjusted copy generation plan, or providing the prompt words of the copy generation plan to the user, receiving a modification instruction from the user for the prompt words of the copy generation plan, modifying the prompt words according to the modification instruction of the prompt words, providing the modified prompt words to the copy generation plan construction model, and obtaining the adjusted copy generation plan output by the copy generation plan construction model based on the modified prompt words; generating the target copy according to the copy generation plan includes: generating the target copy according to the adjusted copy generation plan.

[0008] In one implementation, generating the target copy according to the copy generation plan includes: calling the copy processing model corresponding to each of the subtasks in sequence according to the copy generation plan to generate the target copy.

[0009] In one embodiment, the step of sequentially calling the copy processing model corresponding to each of the subtasks according to the copy generation plan to generate the target copy includes: calling a first processing model corresponding to a first subtask according to the copy type, user demand, target audience and the copy generation plan of the target copy, so as to obtain a first output result according to the copy type, user demand, target audience and the prompt word template corresponding to the first subtask; calling a second processing model corresponding to a second subtask according to the first output result and the copy generation plan, so as to obtain a second output result according to the first output result and the prompt word template corresponding to the second subtask, wherein the second subtask is a subtask located after the first subtask and adjacent to the first subtask in the copy generation plan; and continuing to call the copy processing model corresponding to the subtask located after the second subtask and adjacent to the second subtask according to the second output result and the copy generation plan, until the copy processing model corresponding to each of the subtasks in the copy generation plan has been called to obtain the target copy.

[0010] In one embodiment, the first processing model corresponding to the first subtask is called according to the copy type, user needs, target audience and the copy generation plan of the target copy, so as to obtain the first output result according to the copy type, user needs, target audience and the prompt word template corresponding to the first subtask, including: obtaining the prompt word template corresponding to the first subtask according to the copy generation plan; constructing the first prompt word corresponding to the first processing model corresponding to the first subtask according to the copy type, user needs, target audience and the prompt word template of the target copy; providing the first prompt word to the first processing model, and obtaining the first output result output by the first processing model based on the first prompt word.

[0011] In one embodiment, before calling the second processing model corresponding to the second subtask based on the first output result and the text generation plan, the method also includes: providing the first output result and the first prompt word corresponding to the first processing model to the user; receiving a modification instruction from the user for the first prompt word corresponding to the first processing model; modifying the first prompt word according to the modification instruction to obtain an updated first prompt word; providing the updated first prompt word to the first processing model, and obtaining an updated first output result output by the first processing model according to the updated first prompt word.

[0012] In one embodiment, generating the target copy according to the copy generation plan includes: generating a first copy and quality assessment information of the first copy according to the copy generation plan, the quality assessment information including assessment information for at least one attribute of the copy; determining whether the quality assessment information of the first copy meets a preset condition; and outputting the first copy as the target copy if the quality assessment information of the first copy meets the preset condition.

[0013] In one embodiment, after determining whether the quality assessment information of the first copy meets the preset conditions, the method further includes: if the quality assessment information of the first copy does not meet the preset conditions, determining the copy attributes to be optimized and optimization suggestions corresponding to the first copy according to the quality assessment information; regenerating prompt words for the copy processing model corresponding to the copy attributes to be optimized according to the optimization suggestions to obtain first optimization prompt words; generating a second copy and quality assessment information of the second copy based on the first optimization prompt words and the copy generation plan; using the second copy as the new first copy, jumping to the step of determining whether the quality assessment information of the first copy meets the preset conditions and continuing to execute, and iterating in a loop until the target copy is obtained or the number of loop iterations is greater than a preset threshold.

[0014] In one embodiment, after determining whether the quality assessment information of the first copy meets the preset conditions, the method further includes: if the quality assessment information of the first copy does not meet the preset conditions, providing the quality assessment information of the first copy to the user; receiving a modification instruction sent by the user according to the quality assessment information of the first copy; regenerating prompt words for at least one of the copy processing models according to the modification instruction to obtain a second optimized prompt word; generating a second copy and the quality assessment information of the second copy based on the second optimized prompt word and the copy generation plan; using the second copy as the new first copy, jumping to the step of determining whether the quality assessment information of the first copy meets the preset conditions and continuing to execute, and iterating in a loop until the target copy is obtained or the number of loop iterations is greater than a preset threshold.

[0015] In one implementation, generating the quality assessment information of the first document includes: scoring the first document according to the document requirement description information; and generating the quality assessment information of the first document according to the scoring result.

[0016] In a second aspect, an embodiment of the present invention further provides a copywriting generation device, comprising: an acquisition unit, used to acquire copywriting requirement description information; a construction unit, used to construct a copywriting generation plan according to the copywriting requirement description information, wherein the copywriting generation plan is used to generate a target copywriting; the copywriting generation plan includes a plurality of subtasks executed sequentially, wherein each of the subtasks is used to process a copywriting attribute of the target copywriting through a corresponding copywriting processing model; and a generation unit, used to generate the target copywriting according to the copywriting generation plan.

[0017] In one embodiment, the construction unit includes: a determination module, used to determine the copy type, user needs, and target audience of the target copy based on the copy requirement description information; a construction module, used to construct the copy generation plan based on the copy type, user needs, and target audience of the target copy.

[0018] In one embodiment, the construction module includes: a construction submodule, which is used to construct the prompt words of the copy generation plan according to the copy type, user needs, and target audience of the target copy; and a providing submodule, which is used to provide the prompt words of the copy generation plan to the copy generation plan construction model, and obtain the copy generation plan output by the copy generation plan construction model based on the prompt words.

[0019] In one embodiment, the device also includes: a first providing unit, which is used to provide the copy generation plan to the user after obtaining the copy generation plan output by the copy generation plan construction model based on the prompt word, receive the user's modification instruction for the copy generation plan, and adjust the copy generation plan according to the modification instruction to obtain an adjusted copy generation plan, or provide the prompt word of the copy generation plan to the user, receive the user's modification instruction for the prompt word of the copy generation plan, modify the prompt word according to the modification instruction of the prompt word, provide the modified prompt word to the copy generation plan construction model, and obtain the adjusted copy generation plan output by the copy generation plan construction model based on the modified prompt word; the generating unit is specifically used to generate the target copy according to the adjusted copy generation plan.

[0020] In one implementation, the generating unit is specifically configured to call the document processing model corresponding to each of the subtasks in sequence according to the document generation plan to generate the target document.

[0021] In one embodiment, the generating unit includes: a first calling module, configured to call a first processing model corresponding to a first subtask according to the text type, user demand, target audience and text generation plan of the target text, so as to obtain a first output result according to the text type, user demand, target audience and prompt word template corresponding to the first subtask; a second calling module, configured to call a second processing model corresponding to a second subtask according to the first output result and the text generation plan, so as to obtain a second output result according to the first output result and the prompt word template corresponding to the second subtask, wherein the second subtask is a subtask located after the first subtask and adjacent to the first subtask in the text generation plan; and a third calling module, configured to continue to call the text processing model corresponding to the subtask located after the second subtask and adjacent to the second subtask according to the second output result and the text generation plan, until the text processing model corresponding to each of the subtasks in the text generation plan has been called to obtain the target text.

[0022] In one embodiment, the first calling module is specifically used to: obtain the prompt word template corresponding to the first subtask according to the copy generation plan; construct the first prompt word corresponding to the first processing model corresponding to the first subtask according to the copy type, user needs, target audience and the prompt word template of the target copy; provide the first prompt word to the first processing model, and obtain the first output result of the first processing model based on the first prompt word.

[0023] In one embodiment, the device also includes: a second providing unit, used to provide the first output result and the first prompt word corresponding to the first processing model to the user before calling the second processing model corresponding to the second subtask according to the first output result and the text generation plan; a receiving unit, used to receive the user's modification instruction for the first prompt word corresponding to the first processing model; a first modification unit, used to modify the first prompt word according to the modification instruction to obtain an updated first prompt word; and a third providing unit, used to provide the updated first prompt word to the first processing model, and obtain the updated first output result output by the first processing model according to the updated first prompt word.

[0024] In one embodiment, the generation unit is specifically used to: generate a first copy and quality assessment information of the first copy according to the copy generation plan, the quality assessment information including assessment information for at least one attribute of the copy; determine whether the quality assessment information of the first copy meets a preset condition; and output the first copy as the target copy if the quality assessment information of the first copy meets the preset condition.

[0025] In one embodiment, the generating unit is further configured to: determine, according to the quality evaluation information, the text attributes to be optimized and the optimization suggestions corresponding to the first text when the quality evaluation information of the first text does not meet the preset conditions; regenerate prompt words for the text processing model corresponding to the text attributes to be optimized according to the optimization suggestions to obtain first optimized prompt words; generate a second text and the quality evaluation information of the second text based on the first optimized prompt words and the text generation plan; and use the second text as the new first text, jump to the step of determining whether the quality evaluation information of the first text meets the preset conditions and continue to execute, and iterate in a loop until the target text is obtained or the number of loop iterations is greater than a preset threshold.

[0026] In one embodiment, the generating unit is further configured to: provide the quality assessment information of the first copy to the user if the quality assessment information of the first copy does not meet the preset conditions; receive a modification instruction sent by the user according to the quality assessment information of the first copy; regenerate prompt words for at least one of the copy processing models according to the modification instruction to obtain second optimized prompt words; generate a second copy and the quality assessment information of the second copy based on the second optimized prompt words and the copy generation plan; use the second copy as the new first copy, jump to the step of determining whether the quality assessment information of the first copy meets the preset conditions and continue to execute, and iterate in a loop until the target copy is obtained or the number of loop iterations is greater than a preset threshold.

[0027] In one implementation, the generating unit is further configured to: score the first document according to the document requirement description information; and generate the quality assessment information of the first document according to the scoring result.

[0028] In a third aspect, an embodiment of the present application provides an electronic device, comprising: a housing, a processor, a memory, a circuit board and a power supply circuit, wherein the circuit board is placed inside the space enclosed by the housing, and the processor and the memory are arranged on the circuit board; a power supply circuit is used to supply power to various circuits or devices of the above-mentioned electronic device; the memory is used to store executable program code; the processor runs a program corresponding to the executable program code by reading the executable program code stored in the memory, so as to execute the method described in any of the aforementioned implementation methods.

[0029] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, which stores one or more programs, and the one or more programs can be executed by one or more processors to implement the method described in any of the aforementioned implementation methods.

[0030] The copywriting generation method, device, electronic device and storage medium provided by the embodiments of the present invention can obtain copywriting requirement description information, construct a copywriting generation plan according to the copywriting requirement description information, and generate a target copywriting according to the copywriting generation plan. Since the copywriting generation plan can include multiple subtasks executed sequentially, and each subtask is used to process a copywriting attribute of the target copywriting through a corresponding copywriting processing model, in this way, a special copywriting processing model can be used to process each copywriting attribute of the target copywriting, and each copywriting processing model can be used more professionally to process the problems it is good at solving, so that the target copywriting can have a greater quality improvement in the corresponding copywriting attributes. Since multiple copywriting attributes of the target copywriting are taken into account, the quality of the generated copywriting can be greatly improved compared to the situation in the prior art where the target copywriting is directly given through a large language model. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0032] Figure 1 A flow chart of a method for generating text provided in an embodiment of the present application; Figure 2A detailed flow chart of a method for generating a document provided in an embodiment of the present application; Figure 3 A schematic diagram of the structure of a text generation device provided in one embodiment of the present application; Figure 4 A schematic diagram of the structure of an electronic device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0033] The embodiments of the present application are described in detail below in conjunction with the accompanying drawings. It should be clear that the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0034] In order to enable those skilled in the art to better understand the technical concept, implementation plan and beneficial effects of the embodiments of the present application, specific embodiments are described in detail below.

[0035] like Figure 1 As shown, an embodiment of the present invention provides a copywriting generation method, which may include: S11, obtaining copywriting requirement description information; In this step, the copywriting requirement description information can be obtained in a variety of ways. For example, in one example, the copywriting requirement description information input by the user can be received, and in another example, the copywriting requirement description information stored in advance can be read.

[0036] Specifically, the copywriting requirement description information may refer to various information used to interact with the artificial intelligence model and describe the copywriting to be generated. The specific form of the copywriting requirement description information is not limited, and it can be in the form of text, video or audio, or a combination of multimedia. The content of the copywriting requirement description information can also be very rich, and can include not only the literal meaning of the text itself, but also information obtained through semantic understanding based on context or tone, scene and other information.

[0037] S12, constructing a copywriting generation plan according to the copywriting requirement description information, wherein the copywriting generation plan is used to generate a target copywriting; the copywriting generation plan includes a plurality of subtasks executed sequentially, wherein each of the subtasks is used to process a copywriting attribute of the target copywriting through a corresponding copywriting processing model; In this step, a copywriting generation plan can be constructed according to the copywriting requirement description information. The copywriting generation plan can refer to how to generate the target copywriting according to the acquired copywriting requirement description information, for example, how many steps are needed to generate the target copywriting, what each step mainly does, what is the purpose of doing so, what kind of input is required, what kind of output is there, what is the relationship between these steps or what is the execution order, etc.

[0038] Each of the above steps can be implemented by a subtask. Specifically, in one embodiment of the present invention, the copywriting generation plan may include a plurality of subtasks executed sequentially, wherein each subtask may be used to process a copywriting attribute of the target copywriting through a corresponding copywriting processing model. The copywriting attributes of a copywriting may include multiple types. For example, in one example, the copywriting attributes may include structure, creativity, subject matter, language, logic, emotion, personalization, etc. The corresponding copywriting attributes of different copywriting types may be different. For example, one or more copywriting attributes of a lawyer's letter type copywriting and a novel type copywriting may be different.

[0039] In this step, each copy attribute can be processed by a special copy processing model. Different copy attributes correspond to different copy processing models. In this way, each copy processing model can be used more professionally to deal with the problems it is good at solving, thereby greatly improving the quality of the target copy in terms of the corresponding copy attributes.

[0040] S13, generating the target text according to the text generation plan.

[0041] After obtaining the copywriting generation plan, in this step, the target copywriting can be generated according to the copywriting generation plan. Specifically, the target copywriting can be generated by executing a series of subtasks in the copywriting generation plan.

[0042] The copywriting generation method provided by the embodiment of the present invention can obtain copywriting requirement description information, construct a copywriting generation plan according to the copywriting requirement description information, and generate a target copywriting according to the copywriting generation plan. Since the copywriting generation plan can include multiple subtasks executed sequentially, and each subtask is used to process a copywriting attribute of the target copywriting through a corresponding copywriting processing model, in this way, a special copywriting processing model can be used to process each copywriting attribute of the target copywriting, and each copywriting processing model can be used more professionally to process the problems it is good at solving, so that the target copywriting can have a greater quality improvement in the corresponding copywriting attributes. Moreover, since multiple copywriting attributes of the target copywriting are taken into account, compared with the situation in the prior art where the target copywriting is directly given through a large language model, the quality of the generated copywriting can be greatly improved.

[0043] Specifically, in one embodiment of the present invention, the copy requirement description information can be obtained in step S11 by receiving user input or by reading a pre-stored file, and then, in step 12, a copy generation plan is constructed according to the copy requirement description information.

[0044] In a specific implementation, step S12 constructs a copy generation plan according to the copy requirement description information, which may include: determining the copy type, user requirements, and target audience of the target copy according to the copy requirement description information; and constructing the copy generation plan according to the copy type, user requirements, and target audience of the target copy. Here, the copy type may refer to the subject matter of the copy, such as press releases, contracts, advertisements, novels, scripts, instructions, letters, speeches, etc. User requirements may refer to various requirements of users for the target copy. The target audience may refer to which people or groups the generated target copy is to be shown to, such as college students, women, new employees, or even users of a certain self-media platform.

[0045] In the embodiment of the present invention, the copy requirement description information may contain relatively rich information, and the copy type, user requirements, target audience information, etc. of the target copy can be extracted from the copy requirement description information. For example, in one example, the copy requirement description information may be "generate marketing copy for new smartphones, the target audience is young technology enthusiasts, focusing on photography, performance and appearance features, preferring fashionable, concise and creative language styles, tending to mid-to-high-end consumption and focusing on cost performance", based on which, it can be determined that the copy type of the target copy is "marketing copy", the user requirement is "generate marketing copy, highlight photography, performance and appearance, for technology enthusiasts aged 20-35, mid-to-high-end consumption and focusing on cost performance", and the target audience is "people aged 20-35, who like technology products, have mid-to-high-end consumption habits, and focus on cost performance".

[0046] After determining the target copy's copy type, user needs, and target audience information, a copy generation plan can be constructed based on this information. Specifically, in one embodiment of the present invention, constructing a copy generation plan based on the target copy's copy type, user needs, and target audience can include: constructing a prompt word, i.e., Prompt, for the copy generation plan based on the target copy's copy type, user needs, and target audience; providing the prompt word of the copy generation plan to a copy generation plan construction model, and obtaining the copy generation plan output by the copy generation plan construction model based on the prompt word.

[0047] In this embodiment, the copywriting plan construction model is an artificial intelligence model for constructing a copywriting plan, such as a large language model such as GPT, which can generate a corresponding copywriting plan according to the prompt words input into the model. In order to enable the copywriting plan construction model to generate a more appropriate copywriting plan, in one embodiment of the present invention, a more appropriate prompt word can be constructed according to the copywriting type, user needs, target audience and other information of the target copywriting, and after the constructed prompt word is sent to the copywriting plan construction model, the copywriting plan construction model can output the result corresponding to the prompt word, that is, the copywriting plan. Still taking the aforementioned marketing copywriting embodiment as an example, according to the copywriting type, user needs, target audience and other information of the target copywriting, the generated copywriting plan can include, for example, the following sequential executions: subtasks for processing the creative architecture of the copywriting, subtasks for language optimization, subtasks for emotional rendering, subtasks for personalized processing, subtasks for quality assessment, etc.

[0048] In an embodiment of the present invention, the generated copy generation plan can be directly used to generate the target copy, or it can be provided to the user before being used to generate the target copy, so that the user can understand the specific content of the copy generation plan, or modify the generated copy generation plan, thereby intervening and participating in the formulation of the copy generation plan.

[0049] Specifically, in one embodiment of the present invention, after obtaining the copy generation plan output by the copy generation plan construction model based on the prompt word, the copy generation method provided by the embodiment of the present invention may also include: providing the copy generation plan to the user, receiving the user's modification instructions for the copy generation plan, and adjusting the copy generation plan according to the modification instructions to obtain an adjusted copy generation plan. That is to say, in this embodiment, the user can see the copy generation plan provided to him / her, and modify and adjust the copy generation plan, so that the adjusted copy generation plan better meets the user's needs. Based on this, step S13 generating the target copy according to the copy generation plan may specifically include: generating the target copy according to the adjusted copy generation plan.

[0050] Optionally, in another embodiment of the present invention, after obtaining the copy generation plan output by the copy generation plan construction model based on the prompt words, the copy generation method provided by the embodiment of the present invention may further include: providing the prompt words of the copy generation plan to the user, receiving the user's modification instruction for the prompt words of the copy generation plan, modifying the prompt words according to the modification instruction of the prompt words, providing the modified prompt words to the copy generation plan construction model, and obtaining the adjusted copy generation plan output by the copy generation plan construction model based on the modified prompt words. That is to say, in this embodiment, the user can be allowed to see the prompt words used by the copy generation plan construction model. If the user is not satisfied with the prompt words, the prompt words can be modified, and the output of the copy generation plan construction model is affected by the modification of the prompt words, so as to obtain the adjusted copy generation plan. Based on this, step S13 generating the target copy according to the copy generation plan can specifically include: generating the target copy according to the adjusted copy generation plan.

[0051] The following describes how to generate the target document according to the document generation plan.

[0052] Specifically, in one embodiment of the present invention, step S13 of generating the target copy according to the copy generation plan may include: calling the copy processing model corresponding to each of the subtasks in sequence according to the copy generation plan to generate the target copy. In this embodiment, the copy processing models corresponding to each subtask may be called in sequence according to the execution order of each subtask in the copy generation plan, wherein, based on the logical relationship between the subtasks, the output of the previous model may be used as the input of the next model, and so on, so that the target copy may be obtained by sequentially processing these copy processing models.

[0053] In one embodiment, according to the copy generation plan, sequentially calling the copy processing model corresponding to each of the subtasks to generate the target copy may include: calling a first processing model corresponding to a first subtask according to the copy type, user demand, target audience of the target copy and the copy generation plan, so as to obtain a first output result according to the copy type, user demand, target audience of the target copy and the prompt word template corresponding to the first subtask; calling a second processing model corresponding to a second subtask according to the first output result and the copy generation plan, so as to obtain a second output result according to the first output result and the prompt word template corresponding to the second subtask, wherein the second subtask is a subtask located after the first subtask and adjacent to the first subtask in the copy generation plan; and continuing to call the copy processing model corresponding to the subtask located after the second subtask and adjacent to the second subtask according to the second output result and the copy generation plan, until the copy processing model corresponding to each of the subtasks in the copy generation plan has been called to obtain the target copy.

[0054] In this embodiment, the first subtask may be the first subtask to be executed in the copywriting generation plan. The first subtask may be used to process any copywriting attribute of the target copywriting. For example, in one example, the first subtask may be used to process the copywriting structure, so as to first grasp the overall situation of the target copywriting from a general direction, which is conducive to improving the overall quality of the target copywriting.

[0055] When the first subtask is executed, its corresponding copy processing model - the first processing model will be called. Since the first subtask is the first subtask to be executed, the input information required by the copy processing model can be relatively basic information such as the copy type, user needs, target audience, etc. of the target copy.

[0056] In an embodiment of the present invention, a pre-set prompt word template corresponding to the first processing model can be provided, and the first processing model can obtain a corresponding first output result according to the text type of the target text, user needs, target audience and the prompt word template. Optionally, the first output result can be different according to the different properties of the text that the first subtask is used to process. For example, in one example, the first subtask can be used to process the text structure, and the first output result can be the text structure. In another example, the first subtask can be used to process emotional rendering, and the first output result can be the text after emotional rendering.

[0057] After obtaining the first output result, according to the copy generation plan, it can be known that the next subtask to be executed is the second subtask, and the first output result can be transmitted to the input end of the second subtask as the input of the second subtask to start executing the second subtask. Specifically, the second processing model corresponding to the second subtask can be called to obtain the second output result according to the first output result and the prompt word template corresponding to the second subtask, and so on, until the copy processing model corresponding to each subtask in the copy generation plan has been called to obtain the target copy.

[0058] Specifically, in one embodiment of the present invention, based on the copy type, user needs, target audience and the copy generation plan of the target copy, calling the first processing model corresponding to the first subtask, so as to obtain the first output result based on the copy type, user needs, target audience and the prompt word template corresponding to the first subtask may include: obtaining the prompt word template corresponding to the first subtask according to the copy generation plan; constructing the first prompt word corresponding to the first processing model corresponding to the first subtask according to the copy type, user needs, target audience and the prompt word template of the target copy; providing the first prompt word to the first processing model, and obtaining the first output result output by the first processing model based on the first prompt word.

[0059] In the embodiment of the present invention, each subtask in the copywriting generation plan may have a corresponding prompt word template, which may include some already written prompt information, and may also include some unwritten information blanks that need to be supplemented. Specifically, the output result of the copywriting processing model corresponding to the previous subtask may be filled into the information blank of the prompt word template corresponding to the current subtask, thereby forming the prompt word of the current subtask, and the copywriting processing model corresponding to the current subtask may obtain the corresponding output result according to the prompt word.

[0060] In particular, since the first subtask is the first subtask to be executed in the copy generation plan, the input of the first subtask can be information such as the copy type, user needs, and target audience of the target copy. By filling the copy type, user needs, and target audience of the target copy into the information blanks of the prompt word template of the first subtask, the first prompt word can be obtained, and the first processing model can obtain the corresponding first output result based on the first prompt word.

[0061] In the aforementioned embodiment, after obtaining the first output result, the first output result can be provided to the second subtask so that the second processing model corresponding to the second subtask can be called according to the first output result and the copy generation plan, but the embodiments of the present invention are not limited to this. In other embodiments of the present invention, after obtaining the first output result, before calling the second processing model corresponding to the second subtask according to the first output result and the copy generation plan, the first output result can also be provided to the user, so that the user can participate more in the copy generation.

[0062] Specifically, in one embodiment of the present invention, before calling the second processing model corresponding to the second subtask according to the first output result and the copy generation plan, the copy generation method provided by the embodiment of the present invention may also include: providing the first output result and the first prompt word corresponding to the first processing model to the user; receiving the user's modification instruction for the first prompt word corresponding to the first processing model; modifying the first prompt word according to the modification instruction to obtain an updated first prompt word; providing the updated first prompt word to the first processing model, and obtaining the updated first output result output by the first processing model according to the updated first prompt word.

[0063] That is to say, in this embodiment, the first prompt word and the first output result of the first processing model can be provided to the user, and the user can evaluate whether such first prompt word and such first output result meet his expectations based on his own experience. If not, the user can modify the first prompt word to obtain an updated first prompt word. After providing the updated first prompt word to the first processing model, the first processing model can output the updated first output result, and the updated first output result can also be provided to the user. If the user thinks that it meets his expectations, the updated first output result can be provided to the second subtask to execute the second subtask. If the user thinks that the updated first output result still does not meet his expectations, the updated first prompt word can continue to be modified, and the cycle iterates until it meets the user's expectations.

[0064] In the above-mentioned embodiment, in step S13, the target copy can be generated according to the copy generation plan, but the embodiment of the present invention is not limited thereto. In other embodiments of the present invention, the quality of the generated target copy can also be evaluated, so that on the one hand, the quality of the generated target copy can be more effectively controlled, and on the other hand, it is convenient for the user to have an in-depth understanding of the target copy.

[0065] Specifically, in one embodiment of the present invention, generating the target copy according to the copy generation plan may include: generating a first copy and quality assessment information of the first copy according to the copy generation plan, the quality assessment information including assessment information for at least one attribute of the copy; determining whether the quality assessment information of the first copy meets a preset condition; and outputting the first copy as the target copy if the quality assessment information of the first copy meets the preset condition.

[0066] That is to say, in this embodiment, in addition to the subtasks for generating the target copy, the copy generation plan may also include a quality assessment subtask, which can be used to perform a quality assessment on the generated first copy to obtain corresponding quality assessment information, thereby determining whether the first copy can be output as a qualified target copy.

[0067] Optionally, in an embodiment of the present invention, the specific method of performing quality assessment on the first copy is not limited, as long as it can reflect the quality of the first copy. Exemplarily, in one example, generating quality assessment information of the first copy may include: scoring the first copy according to the copy requirement description information; and generating the quality assessment information of the first copy according to the scoring result. That is, in this embodiment, the first copy may be scored according to the degree of conformity with the copy requirement description information, and the quality assessment information of the first copy may be generated according to the scoring result.

[0068] Exemplarily, in one example, the evaluation dimensions of quality assessment may include: fit with original needs (1-10 points), creativity and novelty (1-10 points), accuracy and fluency of language expression (1-10 points), appropriateness of emotional expression (1-10 points), degree of personalization (1-10 points), overall persuasiveness (1-10 points), etc.

[0069] Based on this, determining whether the quality assessment information of the first copy meets the preset conditions may specifically include: determining whether each scoring result is greater than a preset score threshold (eg, 8 points). If so, it can be determined that the quality of the first copy is qualified and can be output as the target copy.

[0070] In another embodiment of the present invention, after determining whether the quality assessment information of the first copy meets the preset conditions, the copy generation method provided by the embodiment of the present invention may further include: when the quality assessment information of the first copy does not meet the preset conditions, determining the copy attributes to be optimized and optimization suggestions corresponding to the first copy according to the quality assessment information; regenerating prompt words for the copy processing model corresponding to the copy attributes to be optimized according to the optimization suggestions, to obtain first optimized prompt words; generating a second copy and quality assessment information of the second copy based on the first optimized prompt words and the copy generation plan; taking the second copy as the new first copy, jumping to the step of determining whether the quality assessment information of the first copy meets the preset conditions and continuing to execute, and iterating in a loop until the target copy is obtained or the number of loop iterations is greater than a preset threshold.

[0071] That is to say, in this embodiment, if the quality assessment information of the first copy does not meet the preset conditions, for example, one or more scoring results are less than the preset score threshold, the first copy can be optimized according to the quality assessment information. For example, the copy attributes corresponding to the items whose scoring results are less than the preset score threshold can be found to obtain the copy attributes to be optimized, and then the copy attributes to be optimized can be optimized according to the optimization suggestions given in the quality assessment information. For example, according to the optimization suggestions, the prompt words can be regenerated for the copy processing model corresponding to the copy attributes to be optimized to obtain the first optimized prompt words; based on the first optimized prompt words and the copy generation plan, the second copy and the quality assessment information of the second copy are generated. Compared with the first copy, the second copy has been correspondingly optimized. Further, the second copy can be used as the new first copy, and the step of determining whether the quality assessment information of the first copy meets the preset conditions can be jumped to continue execution, and the loop iteration is repeated until the target copy is obtained or the number of loop iterations is greater than the preset threshold.

[0072] In this embodiment, when the quality assessment information of the first copy does not meet the preset conditions, the first copy can be automatically optimized, but the embodiments of the present invention are not limited to this. In other embodiments of the present invention, users can also participate in the optimization of the copy.

[0073] Specifically, in one embodiment of the present invention, after determining whether the quality assessment information of the first copy meets the preset conditions, the copy generation method provided by the embodiment of the present invention may further include: if the quality assessment information of the first copy does not meet the preset conditions, providing the quality assessment information of the first copy to the user; receiving a modification instruction sent by the user according to the quality assessment information of the first copy; according to the modification instruction, regenerating prompt words for at least one of the copy processing models to obtain second optimized prompt words; based on the second optimized prompt words and the copy generation plan, generating a second copy and the quality assessment information of the second copy; using the second copy as the new first copy, jumping to the step of determining whether the quality assessment information of the first copy meets the preset conditions and continuing to execute, and iterating in a loop until the target copy is obtained or the number of loop iterations is greater than a preset threshold.

[0074] In one example, the quality assessment information may include a link to a user interface of the prompt word corresponding to each subtask, and the user may be allowed to enter the corresponding prompt word modification interface through the link and issue a modification instruction. According to the user's modification instruction, the prompt word may be regenerated for the text processing model involved in the modification instruction to obtain a second optimized prompt word, and based on the second optimized prompt word and the text generation plan, a second text and quality assessment information of the second text may be generated.

[0075] The second copy is optimized compared to the first copy. Furthermore, the second copy can be used as the new first copy, and the step of jumping to determine whether the quality assessment information of the first copy meets the preset conditions is continued, and the loop is iterated until the target copy is obtained or the number of loop iterations is greater than the preset threshold.

[0076] The text generation method provided by the embodiment of the present invention is described in detail below through a specific embodiment.

[0077] like Figure 2 As shown, the copywriting generation scheme provided by the embodiment of the present invention may include: S201. Obtain copywriting requirement description information.

[0078] For example, "generate marketing copy for new smartphones, with the target audience being young technology enthusiasts who focus on photography, performance and appearance, prefer fashionable, concise and creative language styles, tend to consume mid- to high-end products and pay attention to cost performance." S202, determining the copy type, user needs, and target audience of the target copy based on the copy requirement description information; For example, the copy type is "marketing copy", and the user demand is "generate marketing copy, highlighting photography, performance and appearance, targeting technology enthusiasts aged 20-35, mid-to-high-end consumption and focus on cost-effectiveness", and the target audience is "people aged 20-35, who are interested in technology products, have mid-to-high-end consumption habits, and focus on cost-effectiveness".

[0079] S203: Construct prompt words for a copy generation plan according to the copy type, user needs, and target audience of the target copy.

[0080] For example, the prompt word can be: "Please construct the copywriting generation plan JSON data based on the following task information. Task information: { "user_demand_description": "Generate marketing copy, highlighting photography, performance and appearance, targeting technology enthusiasts aged 20-35, mid-to-high-end consumers who value cost-effectiveness", "target_audience_info": "{\"age_range\": \"20 - 35 years old\",\"hobbies\": \"Technology products\",\"consumption_habit\": \"Mid- to high-end consumption, focusing on cost-effectiveness\"}", "copywriting_type": "Marketing copywriting" }” S204: providing the prompt words of the text generation plan to the text generation plan construction model, and obtaining the text generation plan output by the text generation plan construction model based on the prompt words.

[0081] For example, the copywriting generation plan obtained can be: ```json { "sub_tasks": [ { "task_name": "Idea Generation", "input_requirements": { "user_demand": "Generate marketing copy, highlighting photography, performance and appearance, targeting technology enthusiasts aged 20-35, mid-to-high-end consumers who value cost-effectiveness", "target_audience": "{\"age_range\": \"20 - 35 years old\",\"hobbies\": \"Technology products\",\"consumption_habit\": \"Mid- to high-end consumption, focusing on cost-effectiveness\"}" }, "expected_output": "Creative framework for copywriting" }, { "task_name": "Language Optimization", "input_requirements": { "initial_copywriting_framework": "[The creative copywriting framework generated by the creative agent]" }, "expected_output": "Text after language optimization" }, { "task_name": "Emotion Adjustment", "input_requirements": { "optimized_copywriting": "[Copywriting output by language optimized agent]", "target_emotional_tone": "[target emotional tone]" }, "expected_output": "Text after sentiment adjustment" }, { "task_name": "Personalization", "input_requirements": { "emotion_adjusted_copywriting": "[Copywriting output by the emotional agent]", "detailed_user_profile": "[Detailed user profile information in JSON format]" }, "expected_output": "Personalized copy" }, { "task_name": "Quality Assessment", "input_requirements": { "personalized_copywriting": "[Personalized copywriting output by the agent]", "original_task_requirement": "Generate marketing copy, highlighting photography, performance and appearance, targeting technology enthusiasts aged 20-35, mid-to-high-end consumers who value cost-effectiveness" }, "expected_output": { "quality_assessment_report": "Quality Assessment Report", "modification_suggestions": "Modification suggestions" } } ] } ``` S205: According to the text generation plan, obtain a prompt word template corresponding to the first subtask.

[0082] For example, the prompt word template corresponding to the first subtask can be: “Based on the following information, generate an innovative copywriting framework. Task description: ... Please provide: copywriting theme, 3 - 5 key selling points, suggested copywriting structure, creative highlights. Make sure the creative idea not only meets user needs, but also is novel and attractive.” S206, constructing a first prompt word corresponding to a first processing model corresponding to a first subtask according to the text type of the target text, user needs, target audience, and prompt word template; For example, the first prompt word can be: “Based on the following information, generate an innovative copywriting framework. Task description: Generate marketing copy, highlighting photography, performance and appearance, targeting technology enthusiasts aged 20-35, mid-to-high-end consumers who focus on cost-effectiveness. User profile: Young technology enthusiasts, aged 20-35, pursuing cutting-edge fashion technology, paying close attention to photography, performance and appearance, mid-to-high-end consumers who focus on cost-effectiveness. Please provide: copywriting theme, 3-5 key selling points, suggested copywriting structure, and creative highlights. Ensure that the creativity not only meets user needs, but is also novel and attractive.” S207: Provide the first prompt word to the first processing model, and obtain a first output result output by the first processing model based on the first prompt word.

[0083] For example, the first output result may be: Copywriting theme: [Mobile phone brand name] Smartphone - The brilliant intersection of technology and fashion, opening up extraordinary possibilities Key selling points: - Ultra-high-definition main camera, like a time catcher, accurately captures every wonderful moment - Powerful chip, like a powerful engine, driving smoothly without lag, ultimate experience - Fashionable appearance design, just like a personalized fashion medal, showing unique taste and charm Suggested copy structure: Introduce the unique positioning of the phone with an attractive opening, elaborate on the outstanding advantages of photography, performance and appearance in the middle part, and end with an appealing call to action to guide purchase.

[0084] Creative highlights: The creative approach cleverly uses vivid metaphors to visualize the camera function of a mobile phone as a "time catcher", the performance as a "power engine", and the appearance as a "fashion medal", which enhances the appeal and memorability of the copy.

[0085] S208: According to the first output result and the text generation plan, call the second processing model corresponding to the second subtask to obtain the second output result according to the first output result and the prompt word template corresponding to the second subtask.

[0086] For example, the prompt word template corresponding to the second subtask can be: “Please optimize the language of the following copy. Original copy: …Optimization requirements: Improve the fluency and coherence of the language; enhance the accuracy and vividness of the expression; adjust the sentence structure to make the copy more rhythmic; ensure the wording is accurate and avoid redundant expressions. Please provide the optimized copy and briefly explain the main improvements.” The second output result can be: "[Mobile phone brand name] Smartphone - The brilliant intersection of technology and fashion opens up extraordinary possibilities. The ultra-high-definition main camera is like the eyes of a time catcher, capturing every inch of wonderful moments with micron-level precision, perfectly freezing the brilliant time without any regrets. Equipped with a powerful chip, like a built-in surging power engine, it instantly activates the ultimate experience of smooth and non-stuttering, and the stuttering and lag will never disappear between fingertip operations. The fashionable appearance design is like a personalized fashion medal, showing unique taste and charm." S209. Continue to call the copy processing model corresponding to the subtask that is located after the second subtask and adjacent to the second subtask according to the second output result and the copy generation plan, until the copy processing model corresponding to each of the subtasks in the copy generation plan has been called, and the first copy and the quality evaluation information of the first copy are obtained.

[0087] S210: Determine whether the quality assessment information of the first document meets the preset conditions. If yes, execute step S211; if no, execute step S212.

[0088] S211: Output the first text as the target text.

[0089] S212: Determine the copy attributes to be optimized and optimization suggestions corresponding to the first copy according to the quality evaluation information; S213, regenerating prompt words for the text processing model corresponding to the text attribute to be optimized according to the optimization suggestion, to obtain a first optimized prompt word; S214, generating a second copy and quality assessment information of the second copy based on the first optimization prompt word and the copy generation plan; S215: Whether the number of loop iterations is greater than a preset threshold; if not, execute step S216; if yes, execute step S217.

[0090] S216: Use the second text as the new first text and continue with step S210.

[0091] S217, end.

[0092] Correspondingly, in a second aspect, an embodiment of the present invention further provides a text generation device, which can effectively improve the quality of generated text.

[0093] like Figure 3 As shown, the text generation device provided by the embodiment of the present invention may include: An acquisition unit 31 is used to acquire the copywriting requirement description information; A construction unit 32 is used to construct a copywriting generation plan according to the copywriting requirement description information, wherein the copywriting generation plan is used to generate a target copywriting; the copywriting generation plan includes a plurality of subtasks executed in sequence, wherein each of the subtasks is used to process a copywriting attribute of the target copywriting through a corresponding copywriting processing model; The generating unit 33 is used to generate the target text according to the text generation plan.

[0094] The copywriting generation device provided by the embodiment of the present invention can obtain copywriting requirement description information, construct a copywriting generation plan according to the copywriting requirement description information, and generate a target copywriting according to the copywriting generation plan. Since the copywriting generation plan can include multiple subtasks executed sequentially, and each subtask is used to process a copywriting attribute of the target copywriting through a corresponding copywriting processing model, in this way, a special copywriting processing model can be used to process each copywriting attribute of the target copywriting, and each copywriting processing model can be used more professionally to process the problems it is good at solving, so that the target copywriting can have a greater quality improvement in the corresponding copywriting attributes. Since multiple copywriting attributes of the target copywriting are taken into account, the quality of the generated copywriting can be greatly improved compared to the situation in the prior art where the target copywriting is directly given through a large language model.

[0095] In one embodiment, the construction unit 32 may include: A determination module, used to determine the copy type, user needs, and target audience of the target copy according to the copy requirement description information; A construction module is used to construct the copy generation plan according to the copy type, user needs, and target audience of the target copy.

[0096] In one embodiment, the building blocks may include: A construction submodule, used to construct prompt words of the copywriting generation plan according to the copywriting type, user needs, and target audience of the target copywriting; A submodule is provided, which is used to provide the prompt words of the text generation plan to the text generation plan construction model, and obtain the text generation plan output by the text generation plan construction model based on the prompt words.

[0097] In one embodiment, the device further comprises: a first providing unit, configured to, after obtaining the copy generation plan outputted by the copy generation plan construction model based on the prompt word, provide the copy generation plan to the user, receive a modification instruction of the copy generation plan from the user, and adjust the copy generation plan according to the modification instruction to obtain an adjusted copy generation plan, or provide the prompt word of the copy generation plan to the user, receive a modification instruction of the prompt word of the copy generation plan from the user, modify the prompt word according to the modification instruction of the prompt word, provide the modified prompt word to the copy generation plan construction model, and obtain the adjusted copy generation plan outputted by the copy generation plan construction model based on the modified prompt word; The generating unit 33 is specifically configured to generate the target text according to the adjusted text generation plan.

[0098] In one implementation, the generating unit 33 is specifically configured to call the document processing model corresponding to each of the subtasks in sequence according to the document generation plan to generate the target document.

[0099] In one implementation, the generating unit 33 may include: A first calling module is used to call a first processing model corresponding to a first subtask according to the text type of the target text, user needs, target audience and the text generation plan, so as to obtain a first output result according to the text type of the target text, user needs, target audience and the prompt word template corresponding to the first subtask; a second calling module, configured to call a second processing model corresponding to a second subtask according to the first output result and the text generation plan, so as to obtain a second output result according to the first output result and the prompt word template corresponding to the second subtask, wherein the second subtask is a subtask in the text generation plan that is located after the first subtask and adjacent to the first subtask; The third calling module is used to continue to call the copy processing model corresponding to the subtask located after the second subtask and adjacent to the second subtask according to the second output result and the copy generation plan, until the copy processing model corresponding to each of the subtasks in the copy generation plan has been called to obtain the target copy.

[0100] In one implementation, the first calling module is specifically configured to: According to the text generation plan, obtaining a prompt word template corresponding to the first subtask; Constructing a first prompt word corresponding to a first processing model corresponding to the first subtask according to the text type of the target text, user needs, target audience and the prompt word template; The first prompt word is provided to the first processing model, and a first output result output by the first processing model based on the first prompt word is obtained.

[0101] In one embodiment, the device may further include: A second providing unit is used to provide the first output result and a first prompt word corresponding to the first processing model to a user before calling a second processing model corresponding to a second subtask according to the first output result and the text generation plan; A receiving unit, configured to receive a user's instruction to modify the first prompt word corresponding to the first processing model; A first modifying unit, configured to modify the first prompt word according to the modification instruction to obtain an updated first prompt word; The third providing unit is configured to provide the updated first prompt word to the first processing model, and obtain an updated first output result output by the first processing model according to the updated first prompt word.

[0102] In one implementation, the generating unit 33 is specifically configured to: generating a first copy and quality assessment information of the first copy according to the copy generation plan, wherein the quality assessment information includes assessment information for at least one attribute of the copy; Determining whether the quality assessment information of the first copy meets a preset condition; When the quality evaluation information of the first document meets a preset condition, the first document is output as the target document.

[0103] In one implementation, the generating unit 33 is further configured to: When the quality evaluation information of the first copy does not meet the preset condition, determining the copy attributes to be optimized and optimization suggestions corresponding to the first copy according to the quality evaluation information; According to the optimization suggestion, regenerate prompt words for the copy processing model corresponding to the copy attribute to be optimized to obtain a first optimized prompt word; generating a second copy and quality assessment information of the second copy based on the first optimization prompt word and the copy generation plan; The second text is used as the new first text, and the step of determining whether the quality assessment information of the first text meets the preset conditions is jumped to continue execution, and the loop iteration is performed until the target text is obtained or the number of loop iterations is greater than a preset threshold.

[0104] In one implementation, the generating unit 33 is further configured to: If the quality assessment information of the first document does not meet the preset condition, providing the quality assessment information of the first document to the user; receiving a modification instruction sent by a user according to the quality evaluation information of the first copy; Regenerate prompt words for at least one of the copywriting processing models according to the modification instruction to obtain second optimized prompt words; generating a second copy and quality assessment information of the second copy based on the second optimization prompt word and the copy generation plan; The second text is used as the new first text, and the step of determining whether the quality assessment information of the first text meets the preset conditions is jumped to continue execution, and the loop iteration is performed until the target text is obtained or the number of loop iterations is greater than a preset threshold.

[0105] In one implementation, the generating unit 33 is further configured to: According to the description information of the copywriting requirement, scoring the first copywriting; The quality assessment information of the first text is generated according to the scoring result.

[0106] The device of this embodiment can be used to perform Figure 1 The technical solution of the method embodiment shown has similar implementation principles and technical effects, which will not be repeated here.

[0107] Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application is shown in FIG. Figure 4 As shown, it may include: a shell 61, a processor 62, a memory 63, a circuit board 64 and a power supply circuit 65, wherein the circuit board 64 is arranged inside the space enclosed by the shell 61, and the processor 62 and the memory 63 are arranged on the circuit board 64; the power supply circuit 65 is used to supply power to various circuits or devices of the above-mentioned electronic device; the memory 63 is used to store executable program codes; the processor 62 runs the program corresponding to the executable program code by reading the executable program code stored in the memory 63, which is used to execute any one of the model training methods provided in the aforementioned embodiments, or any one of the retrieval service providing methods, thereby also being able to achieve corresponding beneficial technical effects, which have been described in detail in the previous text and will not be repeated here.

[0108] The above electronic devices exist in many forms, including but not limited to: (1) Mobile communication devices: These devices are characterized by their mobile communication functions and their main purpose is to provide voice and data communications. These terminals include: smart phones (such as iPhone), multimedia phones, feature phones, and low-end phones.

[0109] (2) Ultra-mobile personal computer devices: These devices fall into the category of personal computers and have computing and processing capabilities, and generally also have mobile Internet access features. These terminals include: PDAs, MIDs, and UMPC devices, such as the iPad.

[0110] (3) Portable entertainment devices: These devices can display and play multimedia content. They include audio and video players (such as iPods), handheld game consoles, e-books, as well as smart toys and portable car navigation devices.

[0111] (4) Server: A device that provides computing services. The server consists of a processor, hard disk, memory, system bus, etc. The server is similar to a general-purpose computer architecture, but because it needs to provide highly reliable services, it has higher requirements in terms of processing power, stability, reliability, security, scalability, and manageability.

[0112] (5) Other electronic devices with data interaction functions.

[0113] Correspondingly, an embodiment of the present application also provides a computer-readable storage medium, which stores one or more programs, and the one or more programs can be executed by one or more processors to implement any model training method provided in the aforementioned embodiments, or any retrieval service providing method, thereby also being able to achieve the corresponding technical effects, which have been described in detail above and will not be repeated here.

[0114] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0115] Each embodiment in this specification is described in a related manner, and the same or similar parts between the embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0116] In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0117] For the convenience of description, the above device is described by dividing the functions into various units / modules. Of course, when implementing the present application, the functions of each unit / module can be implemented in the same or multiple software and / or hardware.

[0118] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0119] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A copywriting generation method, characterized in that: include: Obtain the description information of the copywriting requirements; Constructing a copywriting generation plan according to the copywriting requirement description information, wherein the copywriting generation plan is used to generate a target copywriting; The copywriting generation plan includes a plurality of subtasks executed sequentially, wherein each of the subtasks is used to process a copywriting attribute of the target copywriting through a corresponding copywriting processing model; The target document is generated according to the document generation plan.

2. The method according to claim 1, characterized in that: The constructing of a copywriting generation plan according to the copywriting requirement description information includes: Determine the copy type, user needs, and target audience of the target copy according to the copy requirement description information; The copy generation plan is constructed according to the copy type, user needs, and target audience of the target copy.

3. The method according to claim 2, characterized in that The step of constructing the copywriting generation plan according to the copywriting type, user needs, and target audience of the target copywriting includes: Constructing prompt words for the copy generation plan according to the copy type, user needs, and target audience of the target copy; The prompt words of the text generation plan are provided to a text generation plan construction model, and the text generation plan output by the text generation plan construction model based on the prompt words is obtained.

4. The method according to claim 2, characterized in that: Generating the target document according to the document generation plan includes: According to the copy generation plan, the copy processing model corresponding to each of the subtasks is called in sequence to generate the target copy.

5. The method according to claim 4, characterized in that The step of sequentially calling the document processing model corresponding to each of the subtasks according to the document generation plan to generate the target document comprises: According to the text type, user needs, target audience and the text generation plan of the target text, a first processing model corresponding to the first subtask is called to obtain a first output result according to the text type, user needs, target audience and the prompt word template corresponding to the first subtask; According to the first output result and the text generation plan, calling a second processing model corresponding to a second subtask to obtain a second output result according to the first output result and the prompt word template corresponding to the second subtask, wherein the second subtask is a subtask in the text generation plan that is located after the first subtask and adjacent to the first subtask; Continue to call the copy processing model corresponding to the subtask located after the second subtask and adjacent to the second subtask according to the second output result and the copy generation plan, until the copy processing model corresponding to each of the subtasks in the copy generation plan has been called to obtain the target copy.

6. The method according to claim 5, characterized in that The first processing model corresponding to the first subtask is called according to the text type, user needs, target audience and the text generation plan of the target text, so as to obtain a first output result according to the text type, user needs, target audience and the prompt word template corresponding to the first subtask, including: According to the text generation plan, obtaining a prompt word template corresponding to the first subtask; Constructing a first prompt word corresponding to a first processing model corresponding to the first subtask according to the text type of the target text, user needs, target audience and the prompt word template; The first prompt word is provided to the first processing model, and a first output result output by the first processing model based on the first prompt word is obtained.

7. The method according to any one of claims 1 to 6, characterized in that Generating the target document according to the document generation plan includes: generating a first copy and quality assessment information of the first copy according to the copy generation plan, wherein the quality assessment information includes assessment information for at least one attribute of the copy; Determining whether the quality assessment information of the first copy meets a preset condition; When the quality evaluation information of the first document meets a preset condition, the first document is output as the target document.

8. A copywriting generation device, characterized in that: include: An acquisition unit, used to acquire the description information of the copywriting requirement; A construction unit, configured to construct a copywriting generation plan according to the copywriting requirement description information, wherein the copywriting generation plan is used to generate a target copywriting; The copywriting generation plan includes a plurality of subtasks executed sequentially, wherein each of the subtasks is used to process a copywriting attribute of the target copywriting through a corresponding copywriting processing model; A generating unit is used to generate the target document according to the document generation plan.

9. An electronic device, characterized in that: The electronic device comprises: a housing, a processor, a memory, a circuit board and a power supply circuit, wherein the circuit board is placed inside the space enclosed by the housing, and the processor and the memory are arranged on the circuit board; the power supply circuit is used to supply power to various circuits or devices of the above-mentioned electronic device; the memory is used to store executable program codes; the processor runs a program corresponding to the executable program code by reading the executable program code stored in the memory, so as to execute the method described in any one of the preceding claims 1-7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the method according to any one of claims 1 to 7.

Images