House type data generation method, electronic equipment, storage medium and program product

Abstract

The invention provides a house type data generation method, electronic equipment, a readable storage medium and a computing program product.The method comprises the steps that firstly, a description text and a demand text of a house with a generated house transformation scheme serve as input of a language model, and a scheme text used for describing the house transformation scheme is obtained through the language model; wherein the description text describes a house type structure before house transformation in a natural language, the demand text is used for describing a transformation demand of a house, and then training the generation model by using training data formed by house type data before transformation, the description text, the demand text, a scheme text and house type data after transformation based on the house; and finally, utilizing the trained generation model to generate a reconstruction scheme of the house with the reconstruction requirement and reconstructed house type data.

Description

Technical Field

[0001] The present disclosure relates to technical fields such as apartment type data processing, and in particular to a method for generating apartment type data, an electronic device, a readable storage medium, and a computer program product. Background Art

[0002] In order to improve the comfort and functionality of the living space, the house can be renovated through apartment modification, so as to solve the original design defects, improve space utilization and meet the changes in residents' lifestyles and personalized needs.

[0003] The current apartment design and renovation requires designers to communicate with customers about the renovation requirements, and then manually design the apartment based on the customer's needs. This process usually needs to be repeated many times, and the process is relatively cumbersome, with low communication and design efficiency. Summary of the Invention

[0004] The present disclosure provides a method for generating apartment type data, an electronic device, a readable storage medium, and a computer program product.

[0005] In a first aspect, the present disclosure proposes a method for generating household type data, comprising: using a description text and a demand text of a house for which a house renovation plan has been generated as input to a language model, obtaining a plan text for describing the renovation plan of the house through the language model, wherein the description text describes the household type structure of the house before renovation in natural language, and the demand text is used to describe the renovation needs of the house; training a generation model using training data formed based on the household type data of the house before renovation, the description text, the demand text, the plan text, and the household type data after renovation; and generating a renovation plan and household type data after renovation for a house with renovation needs using the trained generation model.

[0006] According to some embodiments of the present disclosure, the method of generating the description text includes: for a house for which a house renovation plan has been generated, parsing the house type data before renovation of the house, and obtaining a description text that describes the house type structure before renovation in natural language.

[0007] According to some embodiments of the present disclosure, parsing the apartment structure data of the house before renovation includes: parsing the apartment structure data of the house before renovation to obtain information about functional areas in the house; and generating a description text that describes the apartment structure of the house before renovation in natural language based on the information about the functional areas, the description text including a size description of the functional areas and a description of the positional relationship between the functional areas.

[0008] According to some embodiments of the present disclosure, the apartment type data before renovation and the apartment type data after renovation include size data of the functional areas contained in the house, coordinate data of the wall boundaries of multiple walls, and the inclusion relationship between the functional areas and the walls.

[0009] According to some embodiments of the present disclosure, the apartment type data before the renovation and the apartment type data after the renovation include structural type and position data of the opening structure located in the wall, and the structural type includes at least one of a door, a window, and a door opening; the descriptive text also includes a description of the connectivity relationship between functional areas connected through the opening structure.

[0010] According to some embodiments of the present disclosure, the pre-renovation apartment data and the post-renovation apartment data include location data of items in the house and inclusion relationships between the items and functional areas.

[0011] According to some embodiments of the present disclosure, a method of generating the demand text includes: for a house for which a house renovation plan has been generated, generating a demand text containing renovation demand items based on existing information of the house containing renovation demands.

[0012] According to some embodiments of the present disclosure, a demand text containing renovation requirement items is generated based on the existing information containing renovation requirements of the house, including: inputting the existing information containing renovation requirements of the house and corresponding prompt words into a language model, and obtaining the demand text output by the language model, wherein the demand text includes the renovation requirement items extracted from the existing information.

[0013] According to some embodiments of the present disclosure, the solution text forming the training data is provided with annotations for indicating the quality of the solution text.

[0014] According to some embodiments of the present disclosure, a renovation plan and post-renovation apartment type data for a house with renovation needs are generated using the trained generation model, including: in response to receiving user input from a user terminal, obtaining a target requirement text containing pre-renovation apartment type data of the house and a target requirement text for describing the renovation needs of the house through the user input; parsing the pre-renovation apartment type data obtained through the user input to obtain a target description text that describes the apartment type structure of the house before the renovation in natural language; inputting the target requirement text, the target description text and the pre-renovation apartment type data obtained through the user input into the trained generation model; and providing feedback to the user terminal based on the target plan text and the corresponding post-renovation apartment type data output by the generation model.

[0015] According to some embodiments of the present disclosure, feedback is provided to the user terminal based on the target solution text output by the generation model and the corresponding remodeled apartment type data, including: generating a corresponding remodeled apartment type diagram through the corresponding remodeled apartment type data; and generating feedback information based on the remodeled apartment type diagram and the target solution text output by the generation model and providing feedback to the user terminal.

[0016] A second aspect of the present disclosure proposes an electronic device, comprising: a memory storing execution instructions; and a processor executing the execution instructions stored in the memory, so that the processor executes the method described in any one of the above embodiments.

[0017] A third aspect of the present disclosure provides a readable storage medium, wherein the readable storage medium stores a computer program, and when the computer program is executed by a processor, it is used to implement the method described in any of the above embodiments.

[0018] A fourth aspect of the present disclosure provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it is used to implement the method described in any of the above embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The accompanying drawings illustrate exemplary embodiments of the present disclosure and together with the description serve to explain the principles of the present disclosure. These drawings are included to provide a further understanding of the present disclosure and are incorporated in and constitute a part of this specification.

[0020] Figure 1 A schematic diagram illustrating an application scenario of a method for generating apartment type data according to some embodiments of the present disclosure is shown.

[0021] Figure 2-Figure 7 A schematic diagram of the overall process of a method M100 for generating apartment type data according to some embodiments of the present disclosure is shown.

[0022] Figure 8 It is a schematic block diagram of the structure of a device for generating apartment type data according to an embodiment of the present disclosure.

[0023] Figure 9 1 is a schematic block diagram of the structure of an electronic device 1000 according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0024] The present disclosure will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only intended to explain the relevant content and are not intended to limit the present disclosure. It should also be noted that, for ease of description, only the portions relevant to the present disclosure are shown in the accompanying drawings.

[0025] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in the present disclosure can be combined with each other. The technical solution of the present disclosure will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0026] Unless otherwise stated, the exemplary embodiments / examples shown are to be understood as providing exemplary features of various details of some ways in which the technical concepts of the present disclosure can be implemented in practice. Therefore, unless otherwise stated, the features of the various embodiments / examples may be further combined, separated, interchanged, and / or rearranged without departing from the technical concepts of the present disclosure.

[0027] The terms used herein are for the purpose of describing specific embodiments and are not restrictive. As used herein, unless the context clearly indicates otherwise, the singular forms "a (kind, one)" and "the (the)" are also intended to include plural forms. In addition, when the terms "comprise" and / or "include" and their variations are used in this specification, the description indicates the presence of the stated features, wholes, steps, operations, parts, components and / or their groups, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, parts, components and / or their groups. It should also be noted that, as used herein, the terms "substantially", "approximately" and other similar terms are used as approximate terms and not as degree terms, so that they are used to explain the inherent deviations of the measured values, calculated values ​​and / or provided values ​​that will be recognized by those of ordinary skill in the art.

[0028] Among houses sold with a uniform design, the original design of these houses may no longer be able to meet the diverse living needs of the residents, so the house type needs to be modified so that the spatial layout of the house can meet the needs of users.

[0029] The household renovation process typically involves demand communication, plan design, plan delivery, and plan revisions. Meeting user needs may require multiple rounds of communication, design, and revisions. Designers face a complex and time-consuming workflow, requiring manual communication, the creation of floor plans, and repeated revisions based on the owner's requirements. This results in slow and inefficient communication and design, making it difficult to meet user needs. Furthermore, some owners' requirements are vague, requiring designers to understand and extract them from lengthy descriptions, further slowing work progress.

[0030] To this end, the present disclosure proposes a method for generating apartment type data.

[0031] Figure 1A schematic diagram illustrates an application scenario for a method for generating household type data according to some embodiments of the present disclosure. This application scenario may include a client 10 and a server 20. The server 20 can communicate with the client 10 to send and receive data or instructions. In the present disclosure, each of the client 10 and the server 20 includes at least one processor and at least one memory. The server 20 may be configured with a server.

[0032] For example, the client 10 can be a terminal device held by the user, and the server 20 can be a backend device configured by the service provider of the application, which can be an enterprise. When the user uses the application on the client 10, they can interact with the AI ​​customer service through a dialog box and ask questions about apartment renovation. The server 20 can receive the user's needs sent by the client 10, and determine the house information to be renovated and the renovation requirements description sent by the user. Then, based on these contents, it automatically generates the renovated floor plan and renovation plan description by generating a model and feedback to the user.

[0033] exist Figure 1 The shapes and structures of the client 10 and server 20 shown in the figure should not be construed as limiting the scope of protection of this disclosure. In this disclosure, a "terminal device" can be any type of electronic device, such as a mobile phone, tablet computer, laptop computer, or desktop computer. Furthermore, the server configured with the server 20 can be a physical server or a cloud server, and this disclosure does not limit the type of server.

[0034] Figure 2 FIG. 1 shows a schematic diagram of the overall process of a method M100 for generating apartment type data according to some embodiments of the present disclosure. Figure 2 The method shown includes step S110, step S120 and step S130. The method can be executed by an electronic device such as a computer.

[0035] S110 uses the description text and requirement text of the house for which the renovation plan has been generated as input to a language model, and uses the language model to obtain a proposal text describing the renovation plan. The description text describes the house's layout before renovation in natural language, and the requirement text describes the renovation requirements.

[0036] In order to automatically generate renovation plans and floor plans that meet the user's renovation needs through the generative model, the generative model must first be trained. This step is used to obtain training data for the generative model. Training data can be obtained from the data of completed apartment renovations. For example, you can obtain case data of completed apartment renovations recorded in the business system. Since the houses in these cases have already completed the floor renovation, this case data can include the house floor data before the renovation, the house floor description, and the renovation requirements submitted by the user to the service provider.

[0037] Renovation requests can be documented by users through an online app or in-store consultations. These requests describe the renovation requirements for a house requiring renovation. Below is an example of a previously recorded request q1 for house h1: "We live in a house in District Y, X City. We are a family of five, three generations living together: a couple, a five-year-old child, and two elderly parents. We want each bedroom to function independently, while also providing ample storage space to accommodate the growing demand for daily necessities. We also want each partition to be easily accessible."

[0038] The description text can also be provided by the user through an online app (application) or during in-store consultations, and recorded by the service provider. Alternatively, the service provider can generate a natural language description of the house layout using its own parsing tools after obtaining the latest apartment data. The following is a fragment of description text w1 for apartment data d1 generated by the parsing tool: "The apartment information is as follows: ..., and the room information is as follows: Bedroom A information is as follows: Bedroom A has a floor area of ​​..., and is adjacent to the hallway on its right, ...; Bathroom B information is as follows: Bathroom B has a floor area of ​​..., and is adjacent to the hallway on its right, ...."

[0039] The language model used in step S110 can be a generative pre-trained transformer, namely a GPT model, specifically the GPT-4o model. The GPT-4o model can reason about audio, visual, and text in real time. The prompt words used can be: {description text} + {requirement text} + "The above is detailed information interpreted based on a plan view of a two-dimensional floor plan. Based on the description of the above floor plan, please provide a solution that can address these requirements."

[0040] The resulting proposal, e1, contains a renovation plan for house h1 that meets the user's requirements. The following is an example of the proposed proposal e1 generated by the generated model: "The original dining room and adjacent balcony were connected to create a children's room, transforming the two-bedroom unit into a three-bedroom unit. The master bathroom was separated from the master bedroom, with a dry and wet area separated, and integrated into the public area for convenient use by the entire family."

[0041] S120 , training the generation model using training data formed based on the pre-renovation apartment data, description text, requirement text, plan text, and post-renovation apartment data of the house.

[0042] The pre-renovation floor plan data can be provided by the user to the service provider, or it can be stored in the service provider's database. The user can simply provide the service provider with the house address, and the service provider can retrieve the house's floor plan data from the database based on the house address. If the house has been renovated before, the most recent floor plan data will be used as the pre-renovation floor plan data. The following is a fragment of the apartment data d1 of house h1 obtained from the database: "…, {"id": "wall_1_1", "Coordinates": [{"x":XXX, "y": YYY, "z": 0}, {"x": XXX, "y": YYY, "z": 0}], "Wall Width": XX, "Wall Length": XXXX, "Curved Wall Protrusion Distance": 0, "Type List": ["Exterior Wall"], "Wall Accessory List":[]}, {"id": "wall_1_2",…, "Compartment List": [{"id": "Hallway_1_1", "Compartment Name": "Hallway", "Outline": ["Wall_1_5", "Wall_1_6",…",…] . Where XXX and YYY represent specific values.

[0043] The solution text e1 and the renovated apartment data D1 can also be obtained from the case data of house h1. Specifically, the description text w1, the requirement text q1, and the solution text e1 can all be identified and extracted from the conversation records between the user and the customer service or designer in the case data.

[0044] Generative models can use large language models (LLMs), such as the Llama 3.1 model. Llama 3.1 is a very large language model released by Meta, boasting 405 billion parameters. During training, the Llama 3.1 model can be supervised and fine-tuned using the Megatron architecture. Megatron, also a large language model, improves the efficiency and scalability of large-scale model training by implementing technologies such as tensor model parallelism and pipeline model parallelism.

[0045] During the training process, the generative model inputs the pre-renovation apartment data d1, the requirements document q1, and the description document w1 from the training data. The model's parameters are then adjusted by comparing the output of the generative model with the plan document e1 and the post-renovation apartment data D1. After training the generative model using multiple training data, a trained generative model is obtained. The trained generative model can generate a renovation plan (plan document) and post-renovation apartment data that meet the renovation requirements based on the pre-renovation apartment data, description document, and requirements document. The post-renovation apartment data has the same data format as the pre-renovation apartment data.

[0046] Some parameters of the generative model can be set as follows: the value of the learning rate learning_rate can be 0.000008, the value of the exponential decay rate adam_beta1 that controls the gradient can be 0.9, the value of the exponential decay rate adam_beta2 that controls the squared gradient can be 0.999, the value of weight decay weight_decay can be 0.001, the learning rate scheduling strategy lr_scheduler_type can be the cosine annealing strategy cosine, and the total number of training rounds num_train_epochs can be 2 times.

[0047] S130 , using the trained generative model, generates a renovation plan and post-renovation apartment data for the house requiring renovation.

[0048] After the generative model is put into online use, customers can use the client app to submit a renovation request q2 for a specific property h2. The app then sends q2 to the server, which retrieves the latest unit data for h2 from the backend database as pre-renovation unit data d2. The server then generates a description w2 based on this pre-renovation unit data d2, or directly retrieves h2's description w2 from the backend database. The request q2, pre-renovation unit data d2, and description w2 are then fed into the trained generative model. The model then outputs the renovation plan e2 and post-renovation unit data D2, which are then fed back to the client.

[0049] According to the method for generating household type data proposed in the embodiment of the present disclosure, the trained model can be used to automatically generate household renovation plans for the household renovation needs proposed by customers. There is no need for designers to communicate with customers repeatedly and manually design renovation plans, which significantly improves the efficiency of generating and communicating renovation plans, thereby improving the work efficiency of designers and saving manpower and material resources. In addition, the model can also better understand the needs of customers and can correctly understand and extract needs from the customer's redundant descriptions. The accuracy of the renovation plans generated by the model can reach more than 78%, and the customer satisfaction rate of the survey can reach more than 90%.

[0050] For each house that has completed renovation, the historical data of the renovation design may not be systematically preserved, or it may be difficult to accurately extract or identify the descriptive text from the preserved historical data. For example, the descriptive text exists in numerous conversation records, making it difficult to accurately extract a complete description of the house's layout structure before the renovation.

[0051] Based on this, Figure 3 FIG2 shows a schematic diagram of the overall process of a method M100 for generating apartment type data according to another embodiment of the present disclosure. Figure 3 The method M100 for generating apartment type data may further include step S101. Step S101 is performed before step S110 and is used to generate a description text.

[0052] S101: For a house for which a house renovation plan has been generated, the house type data before the renovation is parsed to obtain a description text describing the house type structure before the renovation in natural language.

[0053] The pre-renovation floor plan data d1 can be stored in the service provider's backend database in a data structure or format that facilitates data conversion. For example, the pre-renovation floor plan data for a house can be a JSON structure that stores the house's floor plan. Therefore, after determining the pre-renovation floor plan data d1 for house h1 from the backend database using the house ID, a rule parsing tool can be used to parse data d1, converting the data in the JSON structure into a natural language description, thereby obtaining a description text w1 describing the pre-renovation floor plan of house h1.

[0054] Figure 4 FIG2 shows a schematic diagram of the overall process of a method M100 for generating apartment type data according to another embodiment of the present disclosure. Figure 4 , step S101 may include step S101a and step S101b.

[0055] S101a: For a house for which a house renovation plan has been generated, the house type data before renovation is analyzed to obtain information about functional areas in the house.

[0056] S101b: Based on the information of the functional areas, a description text is generated in natural language to describe the apartment structure before the renovation, wherein the description text includes a description of the dimensions of the functional areas and a description of the positional relationship between the functional areas.

[0057] Functional areas refer to different spaces in a house divided by walls. A functional area can be a bedroom, bathroom, foyer, kitchen, balcony, corridor or other types of areas. From the example of the description text w1 given above, it can be seen that the description text w1 includes the wall position data of multiple walls in the house h1, and also includes a partition list, wherein the partition list lists the types of functional areas included in the house h1 and the number of each type of functional area, and the partition list also contains the wall ID included in each functional area. Therefore, the information on the functional areas in the house obtained after parsing can include the number and area of ​​each functional area and the positional relationship between each functional area. The information on the functional areas in the house obtained is then expressed in the form of natural language to obtain the description text w1.

[0058] The apartment data before and after renovation may include dimension data of functional areas included in the house, coordinate data of the boundaries of multiple walls, and inclusion relationships between functional areas and walls.

[0059] The dimensional data for a functional area can include its area, as well as its horizontal and vertical lengths. The coordinate data for a wall boundary can include the coordinates of two diagonal corner points of a planar wall. The inclusion relationship between a functional area and a wall can be the IDs of the walls within a functional area. For example, bedroom A includes eight walls, wall_1_5 through wall_1_12, forming a concave-shaped bedroom space.

[0060] The data format of the apartment data D1 after renovation is the same as that of the apartment data d1 before renovation, both of which contain the above-mentioned dimension data, coordinate data and inclusion relationship, but the layout of the house after renovation has changed, so at least one of the dimension data, coordinate data and inclusion relationship will be numerically different from that of the apartment data d1 before renovation.

[0061] The pre- and post-renovation apartment layout data may also include the structural type and location data of opening structures located in the wall, where the structural type includes at least one of a door, a window, and a door opening. Accordingly, the description text may also include a description of the connectivity between functional areas connected by the opening structures.

[0062] For example, the pre-renovation apartment data d1 may also include a list of wall accessories for each wall. Wall accessories refer to structures installed on or dependent on the wall, such as built-in furniture, doors, and windows. Opening structures refer to holes in a wall used for passage and communication, such as doors, windows, or door openings. The following is an example of a wall accessory list for wall_1_5 in the pre-renovation apartment data d1: "…, {"id": "wall_1_5",…,"Wall Accessory List": [{"id": "Entry Single Door_1_2", "Coordinates": {"Starting Point": {"x": XXX, "y": YYY}, "Ending Point": {"x": XXXX,"y": YYYY},…"] where XXX and YYY represent specific values.

[0063] As can be seen from this example, the structural type of the opening structure is a single-door, and its location is determined by the starting and ending point coordinates. The data contained in the wall attachment list can be used to determine the connectivity between different functional areas and describe them in the descriptive text. For example, the descriptive text w1 may include the following fragment: "Bedroom A's information is as follows: ..., Bedroom A is adjacent to the aisle on its right and connected by a single-door, ...; Bedroom C's information is as follows: Bedroom C is adjacent to the aisle above it and connected by a single-door, ...." The renovated apartment data D1 can also include the type and location of the aforementioned opening structure.

[0064] In addition, the pre-renovation and post-renovation apartment layout data may also include the location data of items within the house and the inclusion relationships between items and functional areas. For example, the pre-renovation apartment layout data d1 also includes a furniture list that lists the furniture in house h1, along with its type, location coordinates, length, width, rotation angle, and associated room. The associated room represents the inclusion relationship between the furniture and the room.

[0065] By including the size of each functional area, the relative position relationship between adjacent functional areas, the connectivity relationship between different functional areas and other content in the description text, the generative model can better understand the apartment structure before the renovation, which is beneficial to the training effect of the generative model. After the training is completed, it is also beneficial for the generative model to generate high-quality renovation plans that meet user requirements based on real user needs.

[0066] In the historical data of household renovation design, the requirement text may not be preserved in a systematic manner, just like the description text, or it may be difficult to accurately extract or identify the description text from the preserved historical data. For example, the requirement text also exists in numerous conversation records, making it difficult to accurately extract the complete final version of the user's requirements.

[0067] Based on this, continue to refer to Figure 3 The method M100 for generating apartment type data may further include step S102. Step S102 is executed before step S110 and is used to generate a requirement text. Step S102 and step S101 may be executed asynchronously.

[0068] S102 : For a house for which a house renovation plan has been generated, a requirement text including renovation requirement items is generated based on existing information including renovation requirements of the house.

[0069] Existing information containing renovation requirements can be textual requirements entered by users through an online app, such as the example of requirement text q1 given above. Requirement text q1 is stored in a database to form existing information. When training is required, this existing information can be retrieved, and then one or more renovation requirement items can be extracted from it through language processing. For example, the renovation requirement items obtained from the above requirement text q1 include two items: the first item is "At least three bedrooms are required, each with independent functions", and the second item is "Sufficient storage space is required to meet the growing demand for daily necessities in the future."

[0070] Figure 5 FIG2 shows a schematic diagram of the overall process of a method M100 for generating apartment type data according to another embodiment of the present disclosure. Figure 5 Generating a requirement text containing renovation requirement items based on existing information about the house containing renovation requirements (step S102) may specifically include inputting the existing information about the house containing renovation requirements and corresponding prompt words into a language model, thereby obtaining a requirement text output by the language model. The requirement text includes the renovation requirement items extracted from the existing information.

[0071] The language model used in step S102 can also be GPT-4o. The prompt words used can be: {customer description of household renovation needs} + "The above is the customer's needs for household renovation. Based on the above description, please summarize the customer's needs briefly" to obtain the demand text q1.

[0072] After obtaining the description text w1 in step S101 and the requirement text q1 in step S102 , step S110 may be executed to generate the solution text e1 , thereby obtaining a complete piece of training data for a house h1 .

[0073] Generative models can be trained using supervised fine-tuning (SFT). In supervised fine-tuning, a pre-trained model (generative model) is further trained on a specific task using a labeled dataset. The goal is to adjust model parameters to achieve optimal performance for generating apartment renovation plans. Each piece of training data is considered an SFT data point.

[0074] Therefore, the proposal texts that form the training data can be annotated to indicate the quality of the proposal text. This annotation can be used to indicate the quality of the proposal text. For example, the annotation can be used to indicate the quality level of the renovation proposal, with the higher the quality level, the better the proposal meets the user's needs. Alternatively, the annotation can be used to indicate whether the renovation proposal is correct or incorrect, with a correct proposal meeting the user's needs and an incorrect proposal failing to meet them. Annotation can be performed manually on the proposal text in each piece of training data.

[0075] Figure 6 FIG2 shows a schematic diagram of the overall process of a method M100 for generating apartment type data according to another embodiment of the present disclosure. Figure 6 , using the trained generation model to generate a renovation plan and post-renovation apartment data for a house with renovation requirements (step S130) can specifically include the following steps S131, step S132, step S133 and step S134.

[0076] S131 , in response to receiving a user input sent from a user terminal, obtaining, through the user input, the house type data before renovation and the target requirement text for describing the renovation requirements of the house.

[0077] The user opens the app on the user side (i.e., the client) and enters their apartment renovation request to the AI ​​customer service. After clicking "Send," the app sends the target requirement text q2 containing the requested apartment renovation request to the server. It also sends the address or ID of the house h2 to be renovated, or other new information that can be used to uniquely identify the house. The server directly obtains the target requirement text q2 from the information sent by the user side and, using information such as the address or ID of house h2, extracts the current apartment layout data of house h2 from the database, obtaining the pre-renovation apartment layout data d2.

[0078] S132: Parse the pre-renovation apartment data obtained through user input to obtain a target description text that describes the apartment structure before the renovation in natural language.

[0079] The method for parsing pre-renovation apartment data d2 is similar to the method for parsing pre-renovation apartment data d1 described above. In both cases, functional area information is obtained through parsing using a rule-based parsing tool. Based on this functional area information, a natural language description of the pre-renovation apartment structure of house h2 is generated, resulting in target description text w2. Target description text w2 may include information such as the dimensions and areas of the functional areas, the relative positions between them, and the connectivity between them through wall openings. This information serves as information describing the apartment structure.

[0080] S133, inputting the target requirement text, the target description text and the pre-renovation apartment data obtained through user input into the trained generation model.

[0081] The pre-renovation apartment data d2, target description text w2, target requirement text q2, and corresponding prompt words are input into the generative model. The model outputs the renovation plan text e2 and the post-renovation apartment data D2. The specific prompt words used can be: {pre-renovation apartment data json data d2} + {target description text w2} + {target requirement text q2} + "The above is the apartment structure, apartment description, and renovation requirements. Please make reasonable renovations based on this information, provide renovation ideas and plans, and provide the post-renovation apartment data in json format."

[0082] S134: Feedback is provided to the user terminal based on the target solution text and the corresponding renovated apartment data output by the generation model.

[0083] The transformation plan text e2 and the transformed apartment data D2 output by the generated model are integrated and converted into a data structure that can be displayed in a specified format in a dialog box. For example, the model output can be encapsulated in the format of a house card or other clickable format, and then fed back to the user end. The user end displays this information in a dialog box in the form of a card or other specified format. After the user clicks, they can open a page containing a detailed apartment transformation plan.

[0084] Figure 7 FIG2 shows a schematic diagram of the overall process of a method M100 for generating apartment type data according to another embodiment of the present disclosure. Figure 7 Providing feedback to the user terminal based on the target solution text and the corresponding transformed apartment data output by the generation model (step S134) may specifically include steps S13ra and S134b.

[0085] S134a, generating a corresponding remodeled floor plan based on the corresponding remodeled floor plan data.

[0086] S134b: Generate feedback information based on the renovated floor plan and the target solution text output by the generation model and provide feedback to the user end.

[0087] Since users cannot understand the JSON structure of the house type data, when displaying the house type data in the application, it can be displayed in the form of a house type diagram. Therefore, when displaying the renovated house type data, the user end can first convert the renovated JSON house type data sent by the server into the form of a house type diagram, obtain the renovated house type diagram, and then display it together with the renovation plan text e2 on the page displayed after the user clicks the card. It is understandable that the server end can also convert the renovated JSON house type data output by the generated model into a house type diagram, and then send the renovated house type diagram and the renovation plan text e2 as feedback information to the user end, and the user end can directly display the house type diagram without conversion.

[0088] For example, in step S110, in addition to generating a renovation plan using the GPT-4o model, pre-renovation apartment features can also be generated to characterize the pre-renovation apartment's shortcomings. The specific prompt words used could be: {description text} + {requirement text} + "The above is detailed information interpreted based on a two-dimensional floor plan. Based on the above apartment description, please complete the following tasks: 1. List the shortcomings of this apartment one by one and provide explanations; 2. Provide solutions to address these shortcomings." The resulting apartment features could be "Problem 1: The master bathroom is rarely used by a family of five. Although there are two bathrooms, one is in the master bedroom, making it difficult for the five to use the bathroom during peak hours; Problem 2: A two-bedroom apartment cannot meet the living needs of a family of five. A home with three generations and five people living together requires at least three bedrooms, which the original two-bedroom layout clearly cannot meet." The resulting apartment shortcomings can be omitted from training the generative model; only the renovation plan can be used for training.

[0089] After the generative model training is completed and the generative model is put into use, the output of the generative model obtained in step S133 may also include the characteristics of the apartment type before the renovation and / or the characteristics of the apartment type after the renovation. These characteristics can be used to represent the deep thinking process of the generative model.

[0090] The method for generating household type data provided in this embodiment can directly communicate with users, accurately understand customer needs, and reduce communication costs; it can generate feasible household improvement concepts and plans based on the understanding of the owner's needs; it can perform household improvement operations based on the original household type diagram and combine the household improvement concept, and generate a household type structure diagram after the improvement.

[0091] Based on any of the above embodiments, the present disclosure also provides a device for generating apartment type data. Figure 8 This is a schematic block diagram of the structure of a device for generating household type data according to an embodiment of the present disclosure. Figure 8As shown, the device for generating apartment type data includes: a text generation module 110 , a model training module 120 and a solution generation module 130 .

[0092] The text generation module 110 is configured to use the description text and requirement text of the house for which a renovation plan has been generated as input to a language model, and to generate a proposal text describing the renovation plan through the language model. The description text describes the pre-renovation layout of the house in natural language, while the requirement text describes the renovation requirements.

[0093] The model training module 120 is used to train the generation model using training data formed based on the house type data before renovation, description text, demand text, solution text and house type data after renovation.

[0094] The solution generation module 130 is used to generate a renovation solution and post-renovation apartment data for a house requiring renovation using the trained generation model.

[0095] The above-mentioned device for generating apartment type data may be in the form of computer software, and each module of the above-mentioned device for generating apartment type data may be implemented by a computer software module. The implementation process of the functions and effects of each module in the above-mentioned device is detailed in the implementation process of the corresponding steps in the above-mentioned method, and will not be repeated here.

[0096] The execution subject of the method for generating apartment type data in the specific embodiment of the present disclosure may be a computer, a server or other electronic device.

[0097] Therefore, based on any of the above embodiments, the present disclosure further provides an electronic device, which can execute the method for generating apartment type data of any of the above embodiments described in the present disclosure.

[0098] Figure 9 1 is a block diagram illustrating the structure of an electronic device 1000 according to one embodiment of the present disclosure. The hardware structure of the electronic device 1000 can be implemented using a bus architecture. The bus architecture can include any number of interconnecting buses and bridges, depending on the specific application and overall design constraints of the hardware. The bus 1100 connects various circuits including one or more processors 1200, memory 1300, and / or hardware modules. The bus 1100 can also connect various other circuits 1400, such as peripheral devices, voltage regulators, power management circuits, external antennas, etc.

[0099] Bus 1100 may be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Component Architecture (EISA) bus. Buses can be classified as address buses, data buses, control buses, and the like. For ease of illustration, this figure shows only one connecting line, but this does not imply that there is only one bus or only one type of bus.

[0100] The processor 1200 may be a central processing unit (CPU). The processor 1200 may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or a combination of these chips.

[0101] Memory 1300 can be used as a non-transitory computer-readable storage medium to store non-transitory software programs, non-transitory computer executable programs, and modules, such as program instructions of the computer program in the embodiments of the present disclosure. Processor 1200 implements the method for generating apartment type data by executing the non-transitory software programs, instructions, and modules stored in memory 1300.

[0102] The memory 1300 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and application programs required for at least one function; the data storage area may store data created by the processor 1200. In addition, the memory 1300 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 1300 may optionally include a memory remotely located relative to the processor 1200, and these remote memories may be connected to the processor 1200 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0103] The present disclosure also provides a readable storage medium having a computer program stored therein, which is used to implement the above-mentioned method when the computer program is executed by a processor. "Readable storage medium" can be any device that can contain, store, communicate, propagate or transmit a program for use in an instruction execution system, device or equipment or in combination with these instruction execution systems, devices or equipment. More specific examples of readable storage media include the following: an electrical connection portion with one or more wirings (electronic device), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and editable read-only memory (EPROM or flash memory), an optical fiber device, and a portable read-only memory (CDROM), etc.

[0104] The methods of the present disclosure may be implemented in whole or in part through software, hardware, firmware, or any combination thereof. When implemented using software, they may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer programs or instructions. When the computer programs or instructions are loaded and executed, the processes or functions of the present disclosure are performed in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, a network device, a user equipment, a core network device, an OAM, or other programmable device.

[0105] A computer program or instruction can be stored in a readable storage medium or transferred from one readable storage medium to another. For example, the computer program or instruction can be transferred from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless means. The readable storage medium can be any accessible medium or a data storage device such as a server or data center that integrates one or more accessible media. The accessible medium can be a magnetic medium such as a floppy disk, hard disk, or magnetic tape; an optical medium such as a digital video disk; or a semiconductor medium such as a solid-state drive. The computer-readable storage medium can be a volatile or non-volatile storage medium, or can include both volatile and non-volatile types of storage media.

[0106] Those skilled in the art will appreciate that the embodiments of the present disclosure may be provided as methods, systems, or computer program products. Therefore, the present disclosure may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present disclosure may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0107] The present disclosure is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the present disclosure. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0108] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0109] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 The steps for the function specified in one or more boxes.

[0110] In the description of this specification, the description with reference to the terms "one embodiment / method", "some embodiments / methods", "example", "specific example", or "some examples" means that the specific features, structures, or characteristics described in conjunction with the embodiment / method or example are included in at least one embodiment / method or example of the present disclosure. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment / method or example. Moreover, the specific features, structures, or characteristics described may be combined in a suitable manner in any one or more embodiments / methods or examples. In addition, those skilled in the art may combine and combine different embodiments / methods or examples described in this specification and the features of different embodiments / methods or examples, unless they are contradictory.

[0111] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features being referred to. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. Throughout the present disclosure, "plurality" means at least two, such as two, three, etc., unless otherwise specifically defined.

[0112] Those skilled in the art will appreciate that the above embodiments are merely intended to clearly illustrate the present disclosure and are not intended to limit the scope of the present disclosure. Other changes or modifications may be made based on the above disclosure, and such changes or modifications are still within the scope of the present disclosure.

Claims

1. A method for generating apartment type data, characterized in that: include: The description text and requirement text of the house for which the house renovation plan has been generated are used as inputs of the language model, and the language model is used to obtain a plan text for describing the house renovation plan, wherein the description text describes the house structure before the renovation in natural language, and the requirement text is used to describe the renovation requirements of the house; Training the generative model using training data formed based on the pre-renovation apartment data of the house, the description text, the demand text, the solution text, and the post-renovation apartment data; as well as The trained generation model is used to generate a renovation plan for a house that needs renovation and data on the house type after renovation.

2. The method for generating apartment type data according to claim 1, wherein: Methods for generating the description text include: For a house for which a house renovation plan has been generated, the house type data of the house before the renovation is parsed to obtain a description text that describes the house type structure before the renovation in natural language.

3. The method for generating apartment type data according to claim 2, characterized in that: Analyze the house type data before renovation, including: Analyzing the house layout data before renovation to obtain information about functional areas in the house; and Based on the information of the functional areas, a description text is generated to describe the apartment structure before the house renovation in natural language, wherein the description text includes a size description of the functional areas and a description of the positional relationship between the functional areas.

4. The method for generating apartment type data according to claim 2 or 3, characterized in that: The apartment data before and after the renovation include the structural type and location data of the opening structure located in the wall, and the structural type includes at least one of a door, a window, and a door opening; the description text also includes a description of the connectivity relationship between the functional areas connected through the opening structure.

5. The method for generating apartment type data according to claim 1 or 2, characterized in that: Methods for generating the requirement text include: For a house for which a house renovation plan has been generated, a demand text including renovation demand items is generated based on existing information including renovation demand of the house.

6. The method for generating apartment type data according to claim 1, characterized in that: Using the trained generative model, a renovation plan and post-renovation apartment data for a house requiring renovation are generated, including: In response to receiving user input from the user terminal, obtaining, through the user input, the pre-renovation apartment data of the house and the target requirement text for describing the renovation requirements of the house; Parsing the pre-renovation apartment structure data obtained through the user input to obtain a target description text describing the apartment structure before the renovation in natural language; Inputting the target requirement text, the target description text, and the pre-renovation apartment data obtained through the user input into the trained generation model; and Feedback is provided to the user terminal based on the target solution text and the corresponding renovated apartment data output by the generation model.

7. The method for generating apartment type data according to claim 6, characterized in that: Providing feedback to the user terminal based on the target solution text and corresponding renovated apartment data output by the generation model, including: Generate a corresponding remodeled floor plan based on the corresponding remodeled floor plan data; and Feedback information is generated based on the remodeled floor plan and the target solution text output by the generation model, and is fed back to the user terminal.

8. An electronic device, characterized in that: include: a memory storing execution instructions; as well as A processor, wherein the processor executes the execution instructions stored in the memory, so that the processor executes the method according to any one of claims 1 to 7.

9. A readable storage medium, characterized in that: The readable storage medium stores a computer program, which is used to implement the method according to any one of claims 1 to 7 when executed by a processor.

10. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the computer program is used to implement the method according to any one of claims 1 to 7.

Images