A method for generating graphic information

Abstract

The present invention relates to the field of computer technology, and in particular to a method for generating graphic information, comprising: generating a plurality of text information; generating a plurality of image information corresponding to each text information; dividing each input information into a plurality of fields, obtaining each text information and each image information generated for each input information corresponding to a single field, and obtaining a single group of graphic units; determining whether generation parameters for the single group of graphic units meet preset standards based on the obtained download ratio of the single group of graphic units; when it is determined that the generation parameters for the single group of graphic units do not meet the preset standards, determining a corresponding processing method for the single group of graphic units; when it is determined that the generation parameters for the single group of graphic units meet the preset standards, continuously using the current generation parameters to complete the processing of each input information, thereby improving the generation quality of the graphic information and further improving the generation efficiency of the graphic information.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a method for generating graphic and text information. Background Art

[0002] With the rapid development of mobile Internet and content, information overload is becoming more and more serious. It is particularly important to use advanced natural language processing and deep learning algorithms to generate graphic and text content that conforms to the brand image and user preferences based on user needs and settings.

[0003] The use of artificial intelligence technology has achieved automated, efficient and accurate generation of graphic content, helping users quickly and easily create high-quality graphic content, improving marketing effectiveness and user experience, which is currently urgently needed by users.

[0004] Chinese patent publication number: CN111046096A, discloses a method and device for generating graphic and text structured information, including: identifying the paragraphs of the mixed graphic and text layout document to be processed, obtaining paragraph information, wherein the paragraph information includes the paragraph external frame information used to characterize the range of the paragraph; determining the layout graphic and text element information in the paragraph according to the paragraph external frame information; reading, judging and marking the graphic element category information of the layout graphic and text element information to obtain the layout graphic and text information; analyzing, splitting and rearranging the graphic element objects according to the preset reading order of the layout graphic and text information to obtain the graphic and text information; mapping the graphic and text information into graphic and text structured information. This implementation method realizes the accurate generation of graphic and text structured information, thereby enhancing the adaptability of the document displayed on different devices; it can be seen that the above technical solution has the following problems: it does not take into account the continuous monitoring of the user's feedback on the generated graphic and text information, and the graphic and text units corresponding to each field are processed in turn, which affects the efficiency of generating graphic and text information. Summary of the invention

[0005] To this end, the present invention provides a method for generating graphic information to overcome the problem in the prior art that the graphic information generation efficiency is affected by not considering continuous monitoring of user feedback on the generated graphic information and processing the graphic units corresponding to each field in a targeted manner.

[0006] To achieve the above object, the present invention provides a method for generating graphic information, comprising:

[0007] S1, obtaining user input information, and generating a number of copywriting information based on the input information through a language training model;

[0008] S2, sequentially generating a number of image information corresponding to each copy information through the image training model, and a single copy information can correspond to multiple image information;

[0009] S3, obtaining each group of graphic and text units, dividing each input information into several fields based on each field keyword corresponding to each input information within a preset time, obtaining each copy information and each image information generated for each input information corresponding to a single field, and obtaining a single group of graphic and text units; if a single input information contains multiple field keywords, the field keyword with the highest frequency of occurrence is used as its corresponding field keyword;

[0010] S4, based on the obtained download ratio of the single group of graphic and text units, determining whether the generation parameters for the single group of graphic and text units meet the preset standards;

[0011] S5, when it is determined that the generation parameters for the single group of graphic units do not meet the preset standard, determining a corresponding processing method for the single group of graphic units;

[0012] S6, when it is determined that the generation parameters for a single group of graphic and text units meet the preset standard, the current generation parameters are continuously used to complete the processing of each input information.

[0013] Further, for a single group of graphic and text units, determining whether generation parameters for the single group of graphic and text units meet preset standards according to the download ratio;

[0014] If the download ratio is less than or equal to the first preset download ratio, the analysis module determines that the generation parameters for the single group of graphic units do not meet the preset standard, and determines the processing method for the generation parameters for the single group of graphic units not meeting the preset standard according to the difference between the second preset download ratio and the download ratio;

[0015] If the download ratio is less than or equal to the second preset download ratio and greater than the first preset download ratio, the analysis module adjusts the first preset download ratio and the second preset download ratio to corresponding values ​​according to the difference between the download ratio of a single group of graphic units and the average value of the download ratios of each group of graphic units; the first preset download ratio is less than the second preset download ratio;

[0016] If the download ratio is greater than the second preset download ratio, the analysis module determines that the generation parameters for the single group of graphic units meet the preset standards, and continues to use the current generation parameters to complete the processing of each input information;

[0017] The download ratio is the ratio of the amount of downloaded image information to the total amount of image information in a single group of image and text units.

[0018] Furthermore, the analysis module records the difference between the download ratio of a single group of graphic units and the average value of the download ratios of all groups of graphic units as a training deviation, and determines a standard adjustment method for the first preset download ratio and the second preset download ratio according to the training deviation, wherein:

[0019] If the training deviation is less than or equal to the first preset training deviation, the analysis module uses the first preset standard adjustment coefficient to adjust the first preset download ratio and the second preset download ratio to corresponding values;

[0020] If the training deviation is less than or equal to the second preset training deviation and greater than the first preset training deviation, the analysis module uses the second preset standard adjustment coefficient to adjust the first preset download ratio and the second preset download ratio to corresponding values; the first preset training deviation is less than the second preset training deviation;

[0021] If the training deviation is greater than the second preset training deviation, the analysis module uses the third preset standard adjustment coefficient to adjust the first preset download ratio and the second preset download ratio to corresponding values.

[0022] Furthermore, based on the adjusted first preset download ratio and the second preset download ratio, it is re-determined whether the generation parameters of the single group of graphic units meet the preset standards. If the download ratio is still less than or equal to the second preset download ratio and greater than the first preset download ratio, the processing method for the generation parameters of the single group of graphic units not meeting the preset standards is determined according to the application difference between the second preset download ratio and the download ratio.

[0023] Furthermore, the analysis module determines a model processing method for the generation parameters of a single group of graphic units that do not meet the preset standards according to the application difference between the second preset download ratio and the download ratio, wherein:

[0024] If the application difference is greater than the second preset application difference, the analysis module adjusts the data volume of the training data of the image training model to a corresponding value according to the image click data for the single group of graphic units within the preset historical time period;

[0025] If the application difference is less than or equal to the second preset application difference and greater than the first preset application difference, the analysis module re-determines the processing method for the generation parameters of the single group of graphic units not meeting the preset standard according to the ratio of the total browsing time for the single group of graphic units within the preset historical time to the preset browsing time;

[0026] If the application difference is less than or equal to the first preset application difference, the text information corresponding to each group of graphic and text units obtained by the analysis module is used as a training sample, and the language generation model is trained according to the training sample.

[0027] Further, the analysis module re-determines the model secondary processing method for the generation parameters of the single group of graphic units that do not meet the preset standards according to the ratio of the total browsing time for the single group of graphic units within the preset historical time to the preset browsing time, wherein:

[0028] If the duration ratio is less than or equal to the preset duration ratio, the analysis module adjusts the amount of training data of the image training model to a corresponding value according to the image click data for a single group of graphic units within the preset historical duration;

[0029] If the duration ratio is greater than the preset duration ratio, the text information corresponding to each group of graphic units obtained by the analysis module is used as a training sample, and the language generation model is trained based on the training sample.

[0030] Furthermore, the image click data is obtained by obtaining the interval duration of each click on each image information in a single group of image and text units within a preset historical duration, and calculating the duration variance of each interval duration to obtain the image click data;

[0031] The analysis module determines an image adjustment method for the training data of the image training model according to the time length variance, wherein:

[0032] If the duration variance is less than or equal to the first preset duration variance, the analysis module uses the first preset image adjustment coefficient to increase the training data for the image training model to a corresponding value;

[0033] If the duration variance is less than or equal to the second preset duration variance and greater than the first preset duration variance, the analysis module uses the second preset image adjustment coefficient to increase the training data for the image training model to a corresponding value;

[0034] If the duration variance is greater than the second preset duration variance, the analysis module uses a third preset image adjustment coefficient to increase the training data for the image training model to a corresponding value.

[0035] Further, obtaining each group of graphic units whose download ratio of the preset reference number is greater than the second preset download ratio, using the text information corresponding to each group of graphic units obtained as training samples, and training the language generation model according to the training samples;

[0036] The analysis module determines the training adjustment method for the preset reference number according to the total number of downloads of a single group of graphic units, wherein:

[0037] If the total number of downloads is less than or equal to the first preset total number of downloads, the analysis module uses the first preset training adjustment coefficient to adjust the preset reference number to a corresponding value;

[0038] If the total number of downloads is less than or equal to the second preset total number of downloads and greater than the first preset total number of downloads, the analysis module uses the second preset training adjustment coefficient to adjust the preset reference number to a corresponding value;

[0039] If the total number of downloads is greater than the second preset total number of downloads, the analysis module uses a third preset training adjustment coefficient to adjust the preset reference number to a corresponding value.

[0040] Compared with the prior art, the beneficial effect of the present invention lies in that the graphic information corresponding to the input information is generated by the training model, the user's feedback on the generated graphic information is continuously monitored, and the graphic units corresponding to each field are processed in turn in a targeted manner. For a single group of graphic units, when the download ratio is less than or equal to the first preset download ratio, the application rate of the graphic information generated according to the input information is too low. In this case, a specific processing method is specifically determined to process each training model; when the download ratio is less than or equal to the second preset download ratio and greater than the first preset download ratio, the analysis module cannot determine whether the current generation parameters for the single group of graphic units meet the preset standard only based on the download ratio. In this case, the deviation between the download ratio of the single group of graphic units and the average value of the download ratios of each group of units is determined, the first preset download ratio and the second preset download ratio are adjusted, and the training deviation is inversely proportional to the first preset download ratio and the second preset download ratio. The download situation of the generated graphic information is continuously monitored, and when it is determined that the generation parameters of the single group of graphic units do not meet the preset standard, it is processed in time, while improving the generation quality of the graphic information, the generation efficiency of the graphic information is further improved.

[0041] Furthermore, a specific processing method is determined for the case where the generation parameters of a single group of image and text units do not meet the preset standards. When the application difference is greater than the second preset application difference, the single group of image and text units corresponding to a single field have generated image information that does not meet the user's input information, and a large number of users have not obtained further information on the text information. In this case, the amount of training data for the image training model is increased to improve the accuracy of the image training model; when the application difference is less than or equal to the first preset application difference, only a small number of users have not downloaded the generated image and text information. In this case, the text information is abnormal. At this time, the language generation model is trained to improve the accuracy of the text information, which effectively improves the degree of fit between the image and text information and the user input information, and further improves the generation efficiency of the image and text information.

[0042] Furthermore, when the specific processing method cannot be determined only based on the application difference, the corresponding processing method is comprehensively determined in combination with the total browsing time for a single group of image and text units within the preset historical time. If the time ratio is less than or equal to the preset time ratio, the generated image and text information is only browsed for a short time, that is, it is determined not to download the image and text information. At this time, the image training model is processed; when the time ratio is greater than the preset time ratio, the user completes browsing the text information. In this case, the image information meets the user's needs, but the user fails to download the generated image and text information because the text information is not consistent with the input information. At this time, the language generation model is trained, which further improves the generation efficiency of the image and text information while improving the accuracy and personalization of the image and text information. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 A flowchart of the steps of the method for generating graphic information according to an embodiment of the present invention;

[0044] Figure 2 The present invention is a flow chart of an embodiment of determining whether generation parameters for a single group of graphic and text units meet preset standards according to a download ratio. DETAILED DESCRIPTION

[0045] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0046] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.

[0047] It should be noted that, in the description of the present invention, terms such as "up", "down", "left", "right", "inside" and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the drawings. This is merely for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation on the present invention.

[0048] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0049] See also Figure 1 as well as Figure 2 As shown, they are respectively a flowchart of the steps of the method for generating graphic information in an embodiment of the present invention, and a flowchart of determining whether the generation parameters for a single group of graphic units meet the preset standards according to the download ratio; a method for generating graphic information in an embodiment of the present invention includes:

[0050] S1, obtaining user input information, and generating a number of copywriting information based on the input information through a language training model;

[0051] S2, sequentially generating a number of image information corresponding to each copy information through the image training model, and a single copy information can correspond to multiple image information;

[0052] S3, obtaining each group of graphic and text units, dividing each input information into several fields based on each field keyword corresponding to each input information within a preset time, obtaining each copy information and each image information generated for each input information corresponding to a single field, and obtaining a single group of graphic and text units; if a single input information contains multiple field keywords, the field keyword with the highest frequency of occurrence is used as its corresponding field keyword;

[0053] S4, based on the obtained download ratio of the single group of graphic and text units, determining whether the generation parameters for the single group of graphic and text units meet the preset standards;

[0054] S5, when it is determined that the generation parameters for the single group of graphic units do not meet the preset standard, determining a corresponding processing method for the single group of graphic units;

[0055] S6, when it is determined that the generation parameters for a single group of graphic and text units meet the preset standard, the current generation parameters are continuously used to complete the processing of each input information.

[0056] Specifically, for a single group of graphic units, determining whether the generation parameters for the single group of graphic units meet the preset standards according to the download ratio;

[0057] If the download ratio is less than or equal to the first preset download ratio, the analysis module determines that the generation parameters for the single group of graphic units do not meet the preset standard, and determines the processing method for the generation parameters for the single group of graphic units not meeting the preset standard according to the difference between the second preset download ratio and the download ratio;

[0058] If the download ratio is less than or equal to the second preset download ratio and greater than the first preset download ratio, the analysis module adjusts the first preset download ratio and the second preset download ratio to corresponding values ​​according to the difference between the download ratio of a single group of graphic units and the average value of the download ratios of each group of graphic units; the first preset download ratio is less than the second preset download ratio;

[0059] If the download ratio is greater than the second preset download ratio, the analysis module determines that the generation parameters for the single group of graphic units meet the preset standards, and continues to use the current generation parameters to complete the processing of each input information;

[0060] The download ratio is the ratio of the amount of downloaded image information to the total amount of image information in a single group of image and text units.

[0061] Generate graphic information corresponding to the input information through the training model, continuously monitor the user's feedback on the generated graphic information, and process the graphic units corresponding to each field in turn in a targeted manner. For a single group of graphic units, when the download ratio is less than or equal to the first preset download ratio, the application rate of the graphic information generated according to the input information is too low. In this case, a specific processing method is specifically determined to process each training model; when the download ratio is less than or equal to the second preset download ratio and greater than the first preset download ratio, the analysis module cannot determine whether the current generation parameters for the single group of graphic units meet the preset standards based only on the download ratio. In this case, determine the deviation between the download ratio of the single group of graphic units and the average value of the download ratios of each group of units, adjust the first preset download ratio and the second preset download ratio, and the training deviation is inversely proportional to the first preset download ratio and the second preset download ratio. Continuously monitor the download status of the generated graphic information, and timely process it when it is determined that the generation parameters of the single group of graphic units do not meet the preset standards. While improving the generation quality of graphic information, the generation efficiency of graphic information is further improved.

[0062] Specifically, the analysis module records the difference between the download ratio of a single group of graphic units and the average value of the download ratios of all groups of graphic units as a training deviation, and determines a standard adjustment method for the first preset download ratio and the second preset download ratio according to the training deviation, wherein:

[0063] If the training deviation is less than or equal to the first preset training deviation, the analysis module uses the first preset standard adjustment coefficient to adjust the first preset download ratio and the second preset download ratio to corresponding values;

[0064] If the training deviation is less than or equal to the second preset training deviation and greater than the first preset training deviation, the analysis module uses the second preset standard adjustment coefficient to adjust the first preset download ratio and the second preset download ratio to corresponding values; the first preset training deviation is less than the second preset training deviation;

[0065] If the training deviation is greater than the second preset training deviation, the analysis module uses the third preset standard adjustment coefficient to adjust the first preset download ratio and the second preset download ratio to corresponding values.

[0066] Specifically, based on the adjusted first preset download ratio and the second preset download ratio, it is re-determined whether the generation parameters of the single group of graphic units meet the preset standards. If the download ratio is still less than or equal to the second preset download ratio and greater than the first preset download ratio, the processing method for the generation parameters of the single group of graphic units not meeting the preset standards is determined according to the application difference between the second preset download ratio and the download ratio.

[0067] Specifically, the analysis module determines a model processing method for a single group of graphic units whose generation parameters do not meet the preset standard according to the application difference between the second preset download ratio and the download ratio, wherein:

[0068] If the application difference is greater than the second preset application difference, the analysis module adjusts the data volume of the training data of the image training model to a corresponding value according to the image click data for the single group of graphic units within the preset historical time period;

[0069] If the application difference is less than or equal to the second preset application difference and greater than the first preset application difference, the analysis module re-determines the processing method for the generation parameters of the single group of graphic units not meeting the preset standard according to the ratio of the total browsing time for the single group of graphic units within the preset historical time to the preset browsing time;

[0070] If the application difference is less than or equal to the first preset application difference, the text information corresponding to each group of graphic and text units obtained by the analysis module is used as a training sample, and the language generation model is trained according to the training sample.

[0071] Determine a specific processing method for a single group of image and text units whose generation parameters do not meet the preset standards. When the application difference is greater than the second preset application difference, the single group of image and text units corresponding to a single field have generated image information that does not meet the user's input information, and a large number of users have not obtained further information on the text information. In this case, the amount of training data for the image training model is increased to improve the accuracy of the image training model. When the application difference is less than or equal to the first preset application difference, only a small number of users have not downloaded the generated image and text information. In this case, the text information is abnormal. At this time, the language generation model is trained to improve the accuracy of the text information. While effectively improving the degree of fit between the image and text information and the user input information, the generation efficiency of the image and text information is further improved.

[0072] Specifically, the analysis module re-determines the secondary processing method for the model whose generation parameters of a single group of graphic units do not meet the preset standards according to the ratio of the total browsing time of a single group of graphic units within the preset historical time to the preset browsing time, wherein:

[0073] If the duration ratio is less than or equal to the preset duration ratio, the analysis module adjusts the amount of training data of the image training model to a corresponding value according to the image click data for a single group of graphic units within the preset historical duration;

[0074] If the duration ratio is greater than the preset duration ratio, the text information corresponding to each group of graphic units obtained by the analysis module is used as a training sample, and the language generation model is trained based on the training sample.

[0075] When the specific processing method cannot be determined only based on the application difference, the corresponding processing method is comprehensively determined in combination with the total browsing time for a single group of image and text units within the preset historical time. If the time ratio is less than or equal to the preset time ratio, the generated image and text information is only browsed for a short time, that is, it is determined not to download the image and text information. At this time, the image training model is processed; when the time ratio is greater than the preset time ratio, the user completes browsing the text information. In this case, the image information meets the user's needs, but the user fails to download the generated image and text information because the text information is not consistent with the input information. At this time, the language generation model is trained to further improve the generation efficiency of the image and text information while improving the accuracy and personalization of the image and text information.

[0076] Specifically, the image click data acquisition process is to obtain the interval duration of each click on each image information in a single group of image and text units within a preset historical duration, and calculate the duration variance of each interval duration to obtain the image click data;

[0077] The analysis module determines an image adjustment method for the training data of the image training model according to the time length variance, wherein:

[0078] If the duration variance is less than or equal to the first preset duration variance, the analysis module uses the first preset image adjustment coefficient to increase the training data for the image training model to a corresponding value;

[0079] If the duration variance is less than or equal to the second preset duration variance and greater than the first preset duration variance, the analysis module uses the second preset image adjustment coefficient to increase the training data for the image training model to a corresponding value;

[0080] If the duration variance is greater than the second preset duration variance, the analysis module uses a third preset image adjustment coefficient to increase the training data for the image training model to a corresponding value.

[0081] Specifically, each group of graphic units whose download ratio of the preset reference number is greater than the second preset download ratio is obtained, and the text information corresponding to each group of graphic units obtained is used as a training sample, and the language generation model is trained according to the training samples;

[0082] The analysis module determines the training adjustment method for the preset reference number according to the total number of downloads of a single group of graphic units, wherein:

[0083] If the total number of downloads is less than or equal to the first preset total number of downloads, the analysis module uses the first preset training adjustment coefficient to adjust the preset reference number to a corresponding value;

[0084] If the total number of downloads is less than or equal to the second preset total number of downloads and greater than the first preset total number of downloads, the analysis module uses the second preset training adjustment coefficient to adjust the preset reference number to a corresponding value;

[0085] If the total number of downloads is greater than the second preset total number of downloads, the analysis module uses a third preset training adjustment coefficient to adjust the preset reference number to a corresponding value.

[0086] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

[0087] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for generating graphic information, characterized in that: include: S1, obtaining user input information, and generating a number of copywriting information based on the input information through a language training model; S2, sequentially generating a number of image information corresponding to each copy information through the image training model, and a single copy information can correspond to multiple image information; S3, obtaining each group of graphic and text units, dividing each input information into several fields based on each field keyword corresponding to each input information within a preset time, obtaining each copy information and each image information generated for each input information corresponding to a single field, and obtaining a single group of graphic and text units; if a single input information contains multiple field keywords, the field keyword with the highest frequency of occurrence is used as its corresponding field keyword; S4, based on the obtained download ratio of the single group of graphic and text units, determining whether the generation parameters for the single group of graphic and text units meet the preset standards; S5, when it is determined that the generation parameters for the single group of graphic units do not meet the preset standard, determining a corresponding processing method for the single group of graphic units; S6, when it is determined that the generation parameters for the single group of graphic units meet the preset standard, continue to use the current generation parameters to complete the processing of each input information; If the download ratio is less than or equal to the first preset download ratio, the analysis module determines that the generation parameters for the single group of graphic units do not meet the preset standard, and determines the processing method for the generation parameters for the single group of graphic units not meeting the preset standard according to the difference between the second preset download ratio and the download ratio; If the download ratio is less than or equal to the second preset download ratio and greater than the first preset download ratio, the analysis module adjusts the first preset download ratio and the second preset download ratio to corresponding values ​​according to the difference between the download ratio of a single group of graphic units and the average value of the download ratios of each group of graphic units; the first preset download ratio is less than the second preset download ratio; If the download ratio is greater than the second preset download ratio, the analysis module determines that the generation parameters for the single group of graphic units meet the preset standards, and continues to use the current generation parameters to complete the processing of each input information; The download ratio is the ratio of the amount of downloaded image information to the total amount of image information in a single group of image and text units.

2. The method for generating graphic information according to claim 1, characterized in that: The analysis module records the difference between the download ratio of a single group of graphic units and the average value of the download ratios of all groups of graphic units as a training deviation, and determines a standard adjustment method for the first preset download ratio and the second preset download ratio according to the training deviation, wherein: If the training deviation is less than or equal to the first preset training deviation, the analysis module uses the first preset standard adjustment coefficient to adjust the first preset download ratio and the second preset download ratio to corresponding values; If the training deviation is less than or equal to the second preset training deviation and greater than the first preset training deviation, the analysis module uses the second preset standard adjustment coefficient to adjust the first preset download ratio and the second preset download ratio to corresponding values; the first preset training deviation is less than the second preset training deviation; If the training deviation is greater than the second preset training deviation, the analysis module uses the third preset standard adjustment coefficient to adjust the first preset download ratio and the second preset download ratio to corresponding values.

3. The method for generating graphic information according to claim 2, characterized in that: Based on the adjusted first preset download ratio and second preset download ratio, it is re-determined whether the generation parameters of the single group of graphic units meet the preset standards. If the download ratio is still less than or equal to the second preset download ratio and greater than the first preset download ratio, the processing method for the generation parameters of the single group of graphic units not meeting the preset standards is determined according to the application difference between the second preset download ratio and the download ratio.

4. The method for generating graphic information according to claim 3, characterized in that: The analysis module determines a model processing method for a single group of graphic units whose generation parameters do not meet the preset standard according to the application difference between the second preset download ratio and the download ratio, wherein: If the application difference is greater than the second preset application difference, the analysis module adjusts the data volume of the training data of the image training model to a corresponding value according to the image click data for the single group of graphic units within the preset historical time period; If the application difference is less than or equal to the second preset application difference and greater than the first preset application difference, the analysis module re-determines the processing method for the generation parameters of the single group of graphic units not meeting the preset standard according to the ratio of the total browsing time for the single group of graphic units within the preset historical time to the preset browsing time; If the application difference is less than or equal to the first preset application difference, the text information corresponding to each group of graphic and text units obtained by the analysis module is used as a training sample, and the language generation model is trained according to the training sample.

5. The method for generating graphic information according to claim 4, characterized in that: The analysis module re-determines the model secondary processing method for the generation parameters of the single group of graphic units that do not meet the preset standards based on the ratio of the total browsing time for the single group of graphic units within the preset historical time to the preset browsing time, wherein: If the duration ratio is less than or equal to the preset duration ratio, the analysis module adjusts the amount of training data of the image training model to a corresponding value according to the image click data for a single group of graphic units within the preset historical duration; If the duration ratio is greater than the preset duration ratio, the text information corresponding to each group of graphic units obtained by the analysis module is used as a training sample, and the language generation model is trained based on the training sample.

6. The method for generating graphic information according to claim 5, characterized in that: The process of acquiring the image click data is to acquire the interval duration of each click on each image information in a single group of image and text units within a preset historical duration, and calculate the duration variance of each interval duration to obtain the image click data; The analysis module determines an image adjustment method for the training data of the image training model according to the time length variance, wherein: If the duration variance is less than or equal to the first preset duration variance, the analysis module uses the first preset image adjustment coefficient to increase the training data for the image training model to a corresponding value; If the duration variance is less than or equal to the second preset duration variance and greater than the first preset duration variance, the analysis module uses the second preset image adjustment coefficient to increase the training data for the image training model to a corresponding value; If the duration variance is greater than the second preset duration variance, the analysis module uses a third preset image adjustment coefficient to increase the training data for the image training model to a corresponding value.

7. The method for generating graphic information according to claim 6, characterized in that: Acquire each group of graphic and text units whose download ratio of the preset reference number is greater than the second preset download ratio, use the text information corresponding to each group of graphic and text units acquired as training samples, and train the language generation model according to the training samples; The analysis module determines the training adjustment method for the preset reference number according to the total number of downloads of a single group of graphic units, wherein: If the total number of downloads is less than or equal to the first preset total number of downloads, the analysis module uses the first preset training adjustment coefficient to adjust the preset reference number to a corresponding value; If the total number of downloads is less than or equal to the second preset total number of downloads and greater than the first preset total number of downloads, the analysis module uses the second preset training adjustment coefficient to adjust the preset reference number to a corresponding value; If the total number of downloads is greater than the second preset total number of downloads, the analysis module uses a third preset training adjustment coefficient to adjust the preset reference number to a corresponding value.

Images