CN-122023586-A - Image generation model training method and device, and image generation method and device

CN122023586ACN 122023586 ACN122023586 ACN 122023586ACN-122023586-A

Abstract

The embodiment of the specification provides an image generation model training method and device, and an image generation method and device, wherein the image generation model training method comprises the steps of obtaining a plurality of page images of a target application program and image description texts of the page images, determining style description texts corresponding to the target application program, constructing training data pairs according to the page images, the image description texts corresponding to the page images and the style description texts, inputting the training data pairs into an initial image generation model, obtaining a predicted image, and training the initial image generation model according to the predicted image and the page images in the training data pairs, so as to obtain a target image generation model. The subsequent generation model based on the target image can directly generate the image according to the image description text and the style description text, so that the time cost of manual design is greatly reduced, and a large number of potential design combinations can be explored based on the model, thereby realizing efficient, intelligent and innovative image generation.

Inventors

FAN PENGYUAN
LIU JINGMING
ZHAO WEI
WU TONGTONG
YANG FAN
ZHANG LIN

Assignees

北京看云软件有限公司

Dates

Publication Date: 20260512
Application Date: 20260214

Claims (11)

1. An image generation model training method, comprising: Acquiring a plurality of page images of a target application program and image description texts of the page images, and determining style description texts corresponding to the target application program, wherein the image description texts describe and generate the page images through a text generation model; Constructing training data pairs according to each page image, the image description text corresponding to each page image and the style description text; inputting the training data pair into an initial image generation model to obtain a predicted image, and training the initial image generation model according to the predicted image and a page image in the training data pair to obtain a target image generation model.
2. The image generation model training method according to claim 1, wherein the acquiring the plurality of page images of the target application program and the image descriptive text of each page image includes: And acquiring a plurality of page images of the target application program, and generating image description texts corresponding to the page images through a text generation model aiming at the page images.
3. The image generation model training method according to claim 1, wherein said generating a model from the training data pair input initial image to obtain a predicted image, comprises: fusing the image description text and the style description text aiming at the page image in the training data pair to obtain a fused text; And inputting the fused text and the page image into the initial image generation model to obtain a predicted image.
4. The image generation model training method according to claim 1, wherein the training the initial image generation model according to the predicted image and the page image in the training data pair to obtain a target image generation model includes: Training the initial image generation model according to the predicted image and the page image in the training data pair by a target parameter adjustment mode to obtain the target image generation model, wherein the target parameter adjustment mode is a mode of adding a target matrix to the initial image generation model and adjusting parameters of the target matrix.
5. The method for training an image generation model according to any one of claims 1 to 4, wherein before determining the style description text corresponding to the target application program, the method further comprises: A plurality of application programs of different visual styles are determined, and unified style description text is determined for a plurality of page images in the application program for any application program.
6. An image generation method, comprising: and acquiring a style description text and an image description text, and inputting the style description text and the image description text into a target image generation model to obtain a target image, wherein the target image generation model is obtained by the method of any one of claims 1-5.
7. An image generation model training apparatus, comprising: The system comprises an acquisition module, a text generation module and a display module, wherein the acquisition module is configured to acquire a plurality of page images of a target application program and image description texts of the page images, and determine style description texts corresponding to the target application program, wherein the image description texts describe and generate the page images through a text generation model; the construction module is configured to construct training data pairs according to each page image, the image description text corresponding to each page image and the style description text; The training module is configured to input the training data pair into an initial image generation model to obtain a predicted image, train the initial image generation model according to the predicted image and the page image in the training data pair, and obtain a target image generation model.
8. An image generating apparatus, comprising: The generating module is configured to acquire style description text and image description text, input the style description text and the image description text into a target image generating model, and acquire a target image, wherein the target image generating model is acquired through the method of any one of claims 1-5.
9. A computing device, comprising: A memory and a processor; the memory is adapted to store a computer program/instruction, the processor being adapted to execute the computer program/instruction, which when executed by the processor performs the steps of the method according to any one of claims 1 to 6.
10. A computer-readable storage medium, characterized in that it stores a computer program/instruction which, when executed by a processor, implements the steps of the method according to any one of claims 1 to 6.
11. A computer program product comprising computer programs/instructions which, when executed by a processor, implement the steps of the method of any of claims 1 to 6.

Description

Image generation model training method and device, and image generation method and device Technical Field Embodiments of the present disclosure relate to the field of computer technology, and in particular, to an image generation model training method, and one or more embodiments of the present disclosure relate to an image generation model training apparatus, an image generation method, an image generation apparatus, a computing device, a computer readable storage medium, and a computer program product. Background In the field of image design, a traditional design method is dominant, and designers use professional design software and tools to perform manual design or complete creation of design manuscript images based on a plurality of standard templates and frames by virtue of their own rich experience. However, the traditional manual design is inefficient, the designer needs to invest a lot of time and effort for drawing sketches and repeatedly modifying, the process is complicated, and at the same time, the imagination and innovation ability of the designer are limited, so that it is difficult to fully explore all the potential design possibilities. Disclosure of Invention In view of this, the present embodiments provide an image generation model training method. One or more embodiments of the present specification relate to an image generation model training apparatus, an image generation method, an image generation apparatus, a computing device, a computer-readable storage medium, and a computer program product, which solve the technical drawbacks in the prior art. According to a first aspect of embodiments of the present specification, there is provided an image generation model training method, including: Acquiring a plurality of page images of a target application program and image description texts of the page images, and determining style description texts corresponding to the target application program, wherein the image description texts describe and generate the page images through a text generation model; Constructing training data pairs according to each page image, the image description text corresponding to each page image and the style description text; inputting the training data pair into an initial image generation model to obtain a predicted image, and training the initial image generation model according to the predicted image and a page image in the training data pair to obtain a target image generation model. According to a second aspect of embodiments of the present specification, there is provided an image generation model training apparatus comprising: The system comprises an acquisition module, a text generation module and a display module, wherein the acquisition module is configured to acquire a plurality of page images of a target application program and image description texts of the page images, and determine style description texts corresponding to the target application program, wherein the image description texts describe and generate the page images through a text generation model; the construction module is configured to construct training data pairs according to each page image, the image description text corresponding to each page image and the style description text; The training module is configured to input the training data pair into an initial image generation model to obtain a predicted image, train the initial image generation model according to the predicted image and the page image in the training data pair, and obtain a target image generation model. According to a third aspect of embodiments of the present specification, there is provided an image generation method including: And acquiring a style description text and an image description text, and inputting the style description text and the image description text into a target image generation model to obtain a target image, wherein the target image generation model is obtained through the image generation model training method. According to a fourth aspect of embodiments of the present specification, there is provided an image generating apparatus comprising: The generating module is configured to acquire a style description text and an image description text, input the style description text and the image description text into a target image generating model, and acquire a target image, wherein the target image generating model is acquired through the image generating model training method. According to a fifth aspect of embodiments of the present specification, there is provided a computing device comprising: A memory and a processor; The memory is used for storing a computer program/instruction, and the processor is used for executing the computer program/instruction, and the computer program/instruction realizes the steps of the image generation model training method and the image generation method when being executed by the processor. According to a sixth aspect of the embodiments of the present s