Scene processing method and device, electronic equipment and computer readable storage medium

Abstract

The invention discloses a scene processing method and device, electronic equipment and a computer readable storage medium. According to the embodiment of the invention, a virtual scene is displayed through a graphical user interface, first object description information is obtained in response to an input operation on the graphical user interface, and the first object description information is used for describing a first target object; generating prompt information on the basis of the first object description information, generating at least one first function call instruction required for constructing the first target object on the basis of the prompt information through a large language model, and calling a corresponding first object generation function on the basis of the at least one first function call instruction, the first target object is generated in the virtual scene, and the efficiency of generating the first target object can be improved.

Description

Technical Field

[0001] This application relates to the technical field of virtual scenarios, and particularly relates to a scenario processing method, apparatus, electronic device, and computer-readable storage medium. Background Art

[0002] Sandbox games are games that allow players to freely explore, build, and adventure in an open world, such as "Minecraft" and "LEGO Infinite".

[0003] When players build structures in sandbox games, it takes a relatively long time and the building efficiency is low. For example, if players are not familiar with the gameplay of sandbox games, they will spend a lot of time getting started with building. Also, for example, some structures are complex and it also takes a fair amount of time to build. Summary of the Invention

[0004] Embodiments of this application provide a scenario processing method, apparatus, electronic device, and computer-readable storage medium, which can reduce the time for constructing objects in a virtual scenario and improve the efficiency of constructing objects in the virtual scenario.

[0005] In a first aspect, embodiments of this application provide a scenario processing method. A virtual scenario is displayed through a graphical user interface. The method includes:

[0006] In response to an input operation on the graphical user interface, obtain first object description information, where the first object description information is used to describe a first target object;

[0007] Based on the first object description information, generate a prompt message, and through a large language model, generate at least one first function call instruction required to construct the first target object based on the prompt message;

[0008] Call a corresponding first object generation function based on the at least one first function call instruction to generate the first target object in the virtual scenario.

[0009] In a second aspect, embodiments of this application further provide a scenario processing apparatus. The apparatus includes:

[0010] An information acquisition module, configured to obtain first object description information in response to an input operation on the graphical user interface, where the first object description information is used to describe a first target object;

[0011] An instruction generation module, configured to generate a prompt message based on the first object description information, and through a large language model, generate at least one first function call instruction required to construct the first target object based on the prompt message;

[0012] An object generation module, configured to call a corresponding first object generation function based on at least one of the above first function call instructions, so as to generate the above first target object in the above virtual scenario.

[0013] In a third aspect, an embodiment of the present application further provides an electronic device, including a memory storing multiple instructions; the above processor loads instructions from the above memory to execute any one of the scenario processing methods provided by the embodiments of the present application.

[0014] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, where the computer-readable storage medium stores multiple instructions, and the above instructions are suitable for being loaded by a processor to execute any one of the scenario processing methods provided by the embodiments of the present application.

[0015] In the embodiments of the present application, in response to an input operation on a graphical user interface, first object description information is obtained, where the first object description information is used to describe a first target object. Based on the first object description information, prompt information is generated, and at least one first function call instruction required to construct the first target object is generated based on the prompt information through a large language model. Based on the at least one first function call instruction, a corresponding first object generation function is called to generate the first target object in a virtual scenario, so that a user can input the first object description information on the graphical user interface to generate the first target object in the virtual scenario, so that even if the user is not familiar with the construction method and / or the constructed object is complex, the first target object can be generated quickly, reducing the time for generating the first target object and improving the efficiency of generating the first target object. Description of the Drawings

[0016] To more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present application. For those skilled in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0017] Figure 1 It is a schematic diagram of an object processing system provided by an embodiment of the present application;

[0018] Figure 2 It is a schematic flowchart of an embodiment of a scenario processing method provided by an embodiment of the present application;

[0019] Figure 3 It is a schematic diagram of another embodiment of the scenario processing method provided by an embodiment of the present application;

[0020] Figure 4 It is a schematic diagram of another embodiment of the scenario processing method provided by an embodiment of the present application;

[0021] Figure 5 It is a schematic structural diagram of a scenario processing device provided by an embodiment of the present application;

[0022] Figure 6 It is a schematic structural diagram of an electronic device provided by an embodiment of the present application. Detailed implementation manners

[0023] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative efforts shall fall within the protection scope of the present application.

[0024] Before explaining the embodiments of the present application in detail, some terms related to the embodiments of the present application will be explained.

[0025] Among them, in the description of the embodiments of the present application, terms such as "first" and "second" may be used in this article to describe various concepts, but unless otherwise specified, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units does not necessarily limit to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0026] The embodiments of the present application provide a scenario processing method, device, electronic device and computer-readable storage medium. Specifically, the scenario processing method of the embodiments of the present application can be executed by an electronic device, where the electronic device can be a device such as a terminal or a server.

[0027] The terminal can be a terminal device such as a smart phone, a tablet computer, a notebook computer, a touch screen, a game console, a personal computer (PC), a personal digital assistant (PDA), etc. The terminal can also include a client, and the client can be a game application client, a browser client carrying a game program or an instant messaging client, etc.

[0028] The server can be an independent physical server, a server cluster or a distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, Content Delivery Network (CDN), and big data and artificial intelligence platforms.

[0029] For example, as Figure 1 shown, this electronic device is described by taking the terminal 10 as an example. The terminal 10 can display a virtual scene through a graphical user interface. In response to an input operation on the graphical user interface, it obtains first object description information, which is used to describe a first target object, and sends the first object description information to the server 20. The server 20 generates a prompt message based on the first object description information, and through a large language model, generates at least one first function call instruction required to construct the first target object based on the prompt message, and calls the corresponding first object generation function based on the at least one first function call instruction to generate the first target object in the virtual scene, and sends the virtual scene including the first target object to the terminal 10 for display.

[0030] The following will be described in detail with reference to the accompanying drawings respectively. It should be noted that the description order of the following embodiments does not limit the preferred order of the embodiments. Although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order from that shown in the drawings.

[0031] In this embodiment, the terminal is taken as an example for description. This embodiment provides a scene processing method. As Figure 2 shown, this scene processing method displays a virtual scene through a graphical user interface. The specific process of this scene processing method can be as follows:

[0032] 201. In response to an input operation on the graphical user interface, obtain first object description information, which is used to describe a first target object.

[0033] Among them, the Graphical User Interface (GUI) is an interface for the terminal to interact with the user. The virtual scene refers to a virtual environment presented through the graphical user interface, which can be a simulation of the real environment or a completely fictional environment. The type of the virtual scene can be set according to the actual situation. For example, the virtual scene can be a game scene or a virtual scenic spot scene, which is not limited in this embodiment.

[0034] The first target object refers to a scene element that can exist in a virtual scene, and its type can be set according to the actual situation. For example, the first target object can be at least one of a virtual building, a virtual plant, and a virtual animal. The virtual building can be, for example, a virtual villa or a virtual pyramid. The virtual plant can be, for example, virtual grass or virtual trees. The virtual animal can be, for example, a virtual pet or a virtual bird. This embodiment does not make any limitations here.

[0035] The first object description information refers to the information used to describe the first target object. Optionally, the first object description information can include at least one of the name of the first target object, the location information of the first target object in the virtual scene, the size information of the first target object, the material information required to construct the first target object, and the color information of the first target object.

[0036] The type of the first object description information can be set according to the actual situation. For example, the first object description information can include at least one string, at least one piece of voice, and / or at least one image. This embodiment does not make any limitations here. When the first object description information includes at least one string, at least one piece of voice, and at least one image, multi-modal input such as text modality, voice modality, and visual modality is implemented, so that the first target object can be generated according to the multi-modal first object description information, improving the accuracy of the generated first target object and making the generated first target object more in line with the user's needs.

[0037] The terminal can obtain an input operation through an input device, and the type of the input device can be set according to the actual situation. For example, the input device can be at least one of a keyboard, a mouse, a touch screen, a camera, and a microphone. This embodiment does not make any limitations here.

[0038] Optionally, the virtual scene displayed through the graphical user interface can be the scene displayed when the client is running. At this time, the input operation is the input operation of an ordinary user (the ordinary user is, for example, a player). Or, the virtual scene displayed through the graphical user interface can be the scene displayed by running an editor. At this time, the input operation can be the input operation of development staff.

[0039] 202. Based on the first object description information, generate a prompt message, and based on the prompt message through a large language model, generate at least one first function call instruction required to construct the first target object.

[0040] Among them, the prompt message can also be called a prompt, which refers to the information input into the large language model. The large language model (LLM) refers to a trained deep learning model, and its type can be set according to the actual situation. For example, the large language model can be ChatGPT or BERT. This embodiment does not make any limitations here.

[0041] The first function call instruction refers to the information required to call the first object generation function, and its form can be determined according to the type of the first object generation function to be called. For example, when the virtual scene is the game scene of "Minecraft" and the first object generation function is a function in the game engine, the function call instruction can be an MC instruction, which is not limited in this embodiment.

[0042] Optionally, the manner of generating the prompt information based on the first object description information can be set according to the actual situation. For example, the first object description information can be converted into the information in the input format required by the large language model to obtain the prompt information, or the prompt information can be obtained after adding supplementary information to the first object description information. The supplementary information refers to the information used to assist the large language model in better constructing the first target object, and its type can be set according to the actual situation. For example, the supplementary information can be at least one of the material information, example information, and first function information required for constructing the first target object, which is not limited in this embodiment.

[0043] In some embodiments, this embodiment further includes:

[0044] Determine the object type to which the first target object belongs based on the first object description information;

[0045] When the object type is the first target type, obtain the material information required for constructing the first target object;

[0046] Generate prompt information based on the first object description information, including:

[0047] Generate prompt information based on the first object description information and the material information.

[0048] Among them, the material information is used to indicate the materials required for constructing the first target object. For example, when the first target object is a pyramid, the material information can be at least one of smooth quartz blocks, gray concrete, and stone brick walls.

[0049] Optionally, multiple object types can be preset, and objects of different object types have different characteristics. For example, a first target type and a second target type can be preset. The first target type refers to the type that can input object parameters. When the object type of the first target object is the first target type, it means that the first target object can be obtained by setting the parameters input into the first object generation function. For example, the first target type can include a first type and a second type. The first type refers to the type that can input partial object parameters. For example, the first type is the pyramid type, and partial object parameters can be, for example, the material for building the pyramid, the height of the pyramid, or the bottom area. The second type refers to the type that can be composed of each sub-object (where the sub-object can also be called an object component). For example, the second type can be the villa type, and the object of the villa type can be composed of a swimming pool sub-object or a room sub-object. The second target type refers to the type to which the complex structure object belongs. Since it takes more time to build the complex structure object or it cannot be generated, the complex structure object can be pre-built in advance. When the object type of the first target object is the second target type, it means that the first target object is a complex structure object, that is, it means that there is a pre-built first target object or there is a pre-built target object similar to the first target object. Optionally, the second target type can include a third type and a fourth type. The third type refers to the type with a pre-built object, and the fourth type refers to the type similar to the pre-built object.

[0050] When the object type of the first target object is the first target type, it means that the first target object is an object that can be obtained by setting the parameters input into the first object generation function. Therefore, the material information required to build the first target object can be obtained, and then based on the first object description information and the material information, a prompt message can be generated so that the prompt message includes the material information.

[0051] Optionally, the method for determining the object type to which the first target object belongs can be set according to the actual situation. For example, through a classifier, based on the first object description information, the object type to which the first target object belongs can be determined. The classifier can be a model trained with a large amount of data to improve the classification accuracy. The classifier can be, for example, a DNN model or a Bert model.

[0052] When determining the object type to which the first target object belongs through a classifier based on the first object description information, the classification prompt message can be generated first based on the first object description information and the classification prompt message template, and then the classification prompt message can be input into the classifier for classification processing to obtain the object type to which the first target object belongs, further improving the accuracy of the determined object type.

[0053] Among them, the classification prompt information template can be set according to the actual situation. Optionally, the classification prompt information template can also include classification example information to make the classification results obtained by the classifier more accurate.

[0054] The classification prompt information can be, for example:

[0055] Now you are a builder in "Minecraft". Given a user's

Construction Instruction

[0056] If the

Construction Instruction

[0057] <|Category|>The above category name

[0058] For example, if the user inputs "Help me plant a field". Output,

[0059] <|Category|>Farmland.

[0060] Among them, "Help me plant a field. Output, <|Category|>Farmland" is the classification example information. All information in the above classification prompt information except the first object description information is the information in the classification prompt template.

[0061] Optionally, feature extraction can be performed on the first object description information, and the material information can be determined based on the extracted features. Optionally, the parameters passed into the first object generation function may be English information. Therefore, after determining the material information, the corresponding English material information of the material information can be determined from the database, and then the prompt information can be generated based on the English material information and the first object description information. Optionally, the corresponding English material information of the material information can be determined from the database through the Retrieval-augmented Generation (RAG) method.

[0062] In this embodiment, based on the first object description information, the object type to which the first target object belongs is determined. When the object type is the first target type, the material information required to construct the first target object is obtained. Based on the first object description information and the material information, a prompt message is generated, such that when the object type of the first target object is the first target type, the prompt message for the first target object includes the material information, reducing the hallucination caused by the lack of material information (when the material information is missing, the large language model may determine some incorrect material information), thereby ensuring the generation effect of the first target object and the stable controllability of generalization, and further making the finally constructed first target object more in line with the user's needs.

[0063] In some embodiments, this embodiment further includes:

[0064] Based on the object type, the first function information required to construct the first target object is determined from the first database;

[0065] Based on the first object description information and the material information, a prompt message is generated, including:

[0066] Based on the first object description information, the material information, and the first function information, a prompt message is generated.

[0067] Among them, the first function information may refer to the function information of a specific first object generation function, or the first function information may also refer to the function information of a general class of object generation functions. For example, the first function information indicates the function information of the object generation function in the game engine. This embodiment does not make a limitation here.

[0068] The first function information may at least include the first function description information, and the first function description information may include at least one of the function name, function description, and return value. Optionally, the first function information may further include parameter description information, and the parameter description information includes at least one of the parameter name, parameter usage, and valid value. For example, when the first target object is a pyramid, the first function information includes the first function description information and the parameter description information, and the first function description information includes the function name and the parameter description information includes the parameter name and parameter usage, the first function information may be "build_pyramid(start_pos, layers, spacing, material), the starting point start_pos of the bottom layer of the pyramid: [x1, y1, z1], the starting point is located at the center of the projection of the pyramid vertex on the bottom layer, layers represents the number of layers of the pyramid, spacing represents the edge spacing of each layer, and material represents the material".

[0069] In this embodiment, based on the object type, the first function information required to construct the first target object is determined from the first database. Based on the first object description information, material information, and the first function information, a prompt message is generated such that the prompt message includes the first function information, so that the large language model can quickly determine the object generation function called by the first function call instruction, thereby improving the efficiency of determining the first function call instruction based on the prompt message.

[0070] In some embodiments, generating a prompt message based on the first object description information includes:

[0071] Performing a refinement process on the first object description information to obtain second object description information;

[0072] Generating a prompt message based on the second object description information.

[0073] Among them, the large language model can be used to perform a refinement process on the first object description information to obtain a second object description. Specifically, a first prompt message template corresponding to the first target object can be preset, and a refinement prompt message is generated based on the first object description information and the first prompt message template, and the refinement prompt message is input into the large language model for refinement processing to obtain the second object description information.

[0074] For example, the first object description information is "XXX", and the first prompt message template is "Now you are a player of 'Minecraft'. You need to improve the details of constructing this object according to the [construction instruction]. You can improve the description from the following aspects:

[0075] 1. What types of bricks are used for this object? How many possible materials are there? You should list all of them.

[0076] 2. How big is this object approximately? You should estimate the length, width, and height. It is better that the length, width, and height do not exceed 20.

[0077] 3. This object needs to include doors, walls, tables, windows, beds, lights, or torches.

[0078] 4. If the requirements in the above 1-3 are inconsistent with the requirements of the user's [construction instruction], the construction requirements of the [construction instruction] should be followed."

[0079] When the first prompt message template is as shown above, the refinement prompt message generated based on the first object description information and the first prompt message template can be as follows:

[0080] "Now you are a player of 'Minecraft'. You need to improve the details of constructing this object according to the [construction instruction]. You can improve the description from the following aspects:

[0081] 1. What types of bricks are used for this object? There may be several materials, and you should list them all.

[0082] 2. How big is this object approximately? You should estimate the length, width, and height. It is better if the length, width, and height do not exceed 20.

[0083] 3. The object needs to include doors, walls, tables, windows, beds, lights, or torches.

[0084] 4. If there is a conflict with the requirements of the user's [Construction Instructions] in the above 1-3, the construction requirements of the [Construction Instructions] should be followed.

[0085] 5. This is the first object description information [Construction Instructions] input by the user: XXX.

[0086] Since the first object description information may not be described sufficiently, when the first object description information is not described sufficiently, the accuracy of the first function call instruction generated by the large language model based on the prompt information is relatively low. Therefore, in this embodiment, the first object description information is refined to obtain the second object description information, and based on the second object description information, prompt information is generated to achieve generating prompt information based on the more fully described second object description information, so that the accuracy of the first function call instruction generated based on the prompt information is higher.

[0087] In some embodiments, generating prompt information based on the first object description information includes:

[0088] Obtaining example information from the second database;

[0089] Generating prompt information based on the first object description information and the example information.

[0090] Among them, the second database and the first database can be the same database or different databases, and this embodiment does not make a limitation here. Example information refers to examples used to illustrate or explain the output function call instructions.

[0091] For example, the example information can be:

[0092] 1. build_pyramid([x1, y1, z1], 12, 1, ['tuff 0']) # Build a 12-layer pyramid made of tuff, with starting coordinates example [a1, b1, c1] and ending coordinates [a2, b2, c2].

[0093] 2. build_pyramid([x2, y2, z2], 8, 1, ['smooth_stone 0']) # Build an 8-layer pyramid made of smooth stone, with starting coordinates example [a3, b3, c3] and ending coordinates [a4, b4, c4].

[0094] Among the above example information, [x1, y1, z1] represents the position of the pyramid, 12 or 8 represents the number of layers of the pyramid, 1 represents the edge spacing of each layer of the pyramid, and ['tuff 0'] and ['smooth_stone 0'] represent the material information.

[0095] In this embodiment, example information is obtained from the second database, and based on the first object description information and the example information, prompt information is generated, so as to enable the large language model to learn and generate function call instructions based on a small amount of example information through the In-Context Learning (ICL) method, enabling the large language model to perform effective reasoning and decision-making in the task of generating function call instructions, and improving the accuracy of the first function call instruction generated by the large language model.

[0096] In some embodiments, in addition to including at least one first function call instruction, the output result of the large language model may further include the thinking process of the large language model. At this time, the prompt information may further include information indicating the output of the thinking process. Optionally, the output format may be specified in the prompt information, and the output of the thinking process may be set in the output format.

[0097] For example, when the first target object is a pyramid, the prompt information may be:

[0098] <|user instruction|>

[0099] XXX.

[0100] <|requirement|>

[0101] You are now a player of "Minecraft" and need to use Bedrock Edition v1.18.33 to call the relevant function instructions to construct a pyramid. The initial position of the pyramid is at

{}

[0102] The output format is as follows:

[0103] <|thinking|>

[0104] Your thinking on constructing the building

[0105] <|output|>

[0106] Output in the order of the xth item as "x.".

[0107] 1. The first instruction executed outputs # your comment, and the comment needs to add the starting coordinates of the command used, such as

a1, b1, c1

a2, b2, c2

[0109] The output of the nth executed instruction, # Your comment, which should include the starting coordinates of the command used, e.g., [a3, b3, c3] and the ending coordinates [a4, b4, c4].

[0110] Block items and names you may use:

[0111] Smooth quartz block: quartz_block 3, gray concrete: concrete 7, oak planks: planks 0, sandstone: sandstone, black stained glass pane: stained_glass 15, smooth quartz stairs: smooth_quartz_stairs 0, lantern: lantern 0, cobblestone wall: cobblestone_wall 7, brown carpet: carpet 12, spruce fence: fence 1, stone stairs: normal_stone_stairs 0, campfire: campfire 0, smooth stone slab: stone_slab0, jungle wood stairs: jungle_stairs 0, cobblestone: cobblestone 0, end rod: end_rod 0, bookshelf: bookshelf 0, jungle planks: wood 3, wood 11, wooden_slab 3, deepslate tile stairs: deepslate_tile_stairs 0, smooth stone: smooth_stone 0, the name of tuff is tuff0.

[0112] Functions that can be used are:

[0113] 1. build_pyramid(start_pos, layers, spacing, material): Function input. Given the starting position of the bottom layer of the building start_pos: [x1, y1, z1], where the starting point is at the center of the projection of the pyramid vertex on the bottom layer; layers: the number of layers of the pyramid; spacing: the edge spacing of each layer; build a pyramid using the given material material[0] (material is also a list). The output of this operation is a list containing multiple function call instructions.

[0114] Output example:

[0115] <|Idea|>- Example

[0116] According to <|Construction Design|>, I need to first check if there are any available functions. If there are, call the function to build the pyramid; if not, build it layer by layer starting from the bottom layer.

[0117] <|Output|>- Example

[0118] 1. build_pyramid([x1, y1, z1], 12, 1, ['tuff 0']) # Build a 12 - layer pyramid made of tuff, with starting coordinates example [a1, b1, c1] and ending coordinates [a2, b2, c2].

[0119] 2. build_pyramid([x2, y2, z2], 8, 1, ['smooth_stone 0']) # Build an 8 - layer pyramid made of smooth stone, with starting coordinates example [a3, b3, c3] and ending coordinates [a4, b4, c4].

[0120] Note:

[0121] 1. You need to output all instructions without omission; use / setblock or / fill to place and fill items, do not use other version - incompatible instructions.

[0122] 2. Add a '#' comment to each instruction and include the starting and ending coordinates of using the command in the comment.

[0123] 3. If there is no indication to use these building elements, these functions can be not used.

[0124] 4. You cannot directly use the original examples for construction. You need to construct according to <|user instruction|> and <|requirements|>.

[0125] In the above prompt information, the user instruction can be the description information of the second object or the first object. The items and names in "Items and names you may use" are material information. The information in "Available functions" is the first function information. "<|Thought|> Your thinking on constructing the building" is used to instruct the large - language model to output the thinking process. The information in "Output example" is example information. The initial position of the above - mentioned pyramid can be the position of the controlled virtual character in the virtual scene. The above - mentioned "<|requirements|>" can be retrieved from the database according to the object type or the name of the first target object.

[0126] In this embodiment, the output result of the large - language model includes at least one first function call instruction and the thinking process of the large - language model, enhancing the user's understanding of the processing process of the large - language model, improving the user's trust in the large - language model, enabling the user to better understand the output result of the large - language model and optimize the large - language model.

[0127] 203. Call the corresponding first object generation function based on at least one first function call instruction to generate the first target object in the virtual scene.

[0128] Among them, the first object generation function refers to a function used to generate a first target object in a virtual scene. In this embodiment, through a large language model, based on the first object description information input by the user, at least one first function call instruction is generated, and the corresponding first object generation function is called based on at least one first function call instruction to generate a first target object in the virtual scene. This enables the user to only input the first object description information to generate the first target object in the virtual scene, improving the efficiency of generating the first target object. For novice players, even if they are not familiar with the construction method, they can quickly get started with the construction, enhancing the game experience. For experienced players, it can save time in constructing complex objects, allowing players to spend more time and energy exploring other gameplay in the virtual scene, increasing the playability of the game, and thus improving the overall fun of the game. When the user is a development staff member, it can also improve the development efficiency and increase the output of constructed objects.

[0129] In some embodiments, this embodiment further includes:

[0130] In response to a first input operation on the graphical user interface, obtain scene editing information;

[0131] Based on the large language model and the scene editing information, edit the virtual scene.

[0132] Among them, the scene editing information may include at least one of object editing information and function editing information. The object editing information is used to edit a second target object in the virtual scene, and the function editing information is used to edit the function code of the second object generation function required to construct the second target object in the virtual scene. Optionally, the scene editing information can be information input by the player or information input by the development staff.

[0133] The second target object can be any one of the first target objects or not the first target object. The object editing information refers to information on how to edit the second target object in the virtual scene, which may include at least one of object modification information and object deletion information. The object modification information can be used to modify at least one of the position, size, and color of the second target object.

[0134] In this embodiment, in response to a first input operation on the graphical user interface, obtain scene editing information, and based on the large language model and the scene editing information, edit the virtual scene, realizing the editing of the virtual scene through the large language model, enabling the user to edit the second target object in the virtual scene and / or the function code corresponding to the virtual scene, improving the flexibility of constructing the virtual scene.

[0135] In some embodiments, the scene editing information includes object editing information, which can be used to modify and / or delete a second target object.

[0136] In some embodiments, the scene editing information includes object editing information, and the object editing information includes object position adjustment information of a second target object in a virtual scene. Editing the virtual scene based on a large language model and the object editing information includes:

[0137] Through the large language model, based on the object position adjustment information, generating at least one second function call instruction required for editing the second target object;

[0138] Based on at least one second function call instruction, calling a corresponding second object generation function to adjust the position of the second target object in the virtual scene.

[0139] Among them, the position adjustment information can be used to indicate the moving direction and moving distance of the second target object. The moving direction can include at least one of up, down, left, right, and symmetric flipping. Generating adjustment prompt information based on the position adjustment information, and inputting the adjustment prompt information into the large language model. The large language model first determines the target position after adjustment of the second target object based on the position adjustment information and the position information of the second target object (where the target position can be an absolute position or a relative position). Based on the target position, generating at least one second function call instruction, and through at least one second function call instruction, calling a corresponding second object generation function to adjust the position of the second target object in the virtual scene.

[0140] Alternatively, the position adjustment information can include the position information and layout mode information of the second target object. Generating adjustment prompt information based on the position adjustment information, and inputting the adjustment prompt information into the large language model. The large language model generates at least one second function call instruction based on the adjustment prompt information, and through at least one second function call instruction, calls a corresponding second object generation function to adjust the position of the second target object in the virtual scene, so as to adjust the layout mode of the second target object.

[0141] Among them, the position information of the second target object can be absolute position information or relative position information. Different layout mode information is used to indicate different layout modes. The layout mode can be, for example, at least one of convex, concave, one-shaped, L-shaped, inverted L-shaped, cross-shaped, and square-shaped. At this time, the second target object can include at least two, and the second object generation function can include a layout function. The types of at least two second target objects can be the same. For example, at least two second target objects are both houses, or the types of at least two second target objects can be different. For example, at least two second target objects can be a house and a fence.

[0142] Optionally, the adjustment prompt information may include not only position adjustment information but also layout example information.

[0143] In this embodiment, the object editing information includes object position adjustment information. Through the large language model, based on the object position adjustment information, at least one second function call instruction required for editing the second target object is generated, and the corresponding second object generation function is called based on the at least one second function call instruction to adjust the position of the second target object in the virtual scene, so as to realize the adjustment of the position of the second target object. When the position of the second target object does not meet the user's requirements, it is not necessary to reconstruct the second target object, and only the second target object needs to be adjusted, improving the efficiency of generating the second target object that meets the user's requirements.

[0144] In some embodiments, the scene editing information includes the function code corresponding to the second object generation function required to construct the second target object. Based on the large language model and the object editing information, the virtual scene is edited, including:

[0145] Through the large language model, the function code is edited to edit the corresponding second object generation function in the virtual scene.

[0146] Among them, the editing of the function code may include at least one of adding function code, modifying function code, and deleting function code. Specifically, the editing prompt information can be generated based on the function code, and the editing prompt information is input into the large language model for editing to obtain the edited function code, thereby realizing the editing of the second object generation function.

[0147] For example, when the second target object is a table and the function code is modified, the editing prompt information may be: You now need to modify the function for building the table according to Bedrock Edition v1.18.33 of Minecraft. There may be problems with the placement of chairs in certain directions in the following function for building the table. You need to check and modify the corresponding code:

[0148] XXXXXXXXXXX;

[0149] return commands.

[0150] For another example, when the second target object is a house and the second object generation function is a function for generating a house, when adding function code, the editing prompt information may be:

[0151] Now you need to modify the function for the house door in Minecraft Bedrock Edition v1.18.33 using magic commands. This is the code for placing the house, and the coordinate system is the same as the reference coordinate system for using commands in Minecraft Bedrock Edition v1.18.33.

[0152] You now need to modify the following function to add houses 5 - 10. House 5 is located to the left of House 1 and has the same size as House 3. House 6 is located below House 1 and has the same size as House 4. House 7 is located to the right of House 1 and has the same size as House 4. The layout of Houses 5 - 7 and the spacing between Houses 1, 3, and 4 are the same, and for Houses 8 - 10.

[0153] Similarly, you need to add the above coordinates to the following function code:

[0154] def place_multiple_houses(pos):

[0155] center_x, center_y, center_z = pos

[0156] # House 1: 10x25, located below, facing the center

[0157] start_pos1 = [center_x - 21, center_y, center_z - 12]

[0158] end_pos1 = [center_x - 12, center_y, center_z + 12]

[0159] # House 2: 10x13, located above, facing the center

[0160] start_pos2 = (center_x + 11, center_y, center_z - 6)

[0161] end_pos2 = (center_x + 20, center_y, center_z + 6)

[0162] # House 3: 12x10, located on the left, facing the center

[0163] start_pos3 = [center_x - 6, center_y, center_z - 15]

[0164] end_pos3 = [center_x + 5, center_y, center_z - 5]

[0165] # House 4: 12x10, located on the right, facing the center

[0166] start_pos4 = (center_x - 6, center_y, center_z + 5)

[0167] end_pos4 = (center_x + 5, center_y, center_z + 15).

[0168] In this embodiment, the scene editing information includes the function code corresponding to the second object generation function required to construct the second target object. Through the large language model, the function code is edited to edit the corresponding second object generation function in the virtual scene, so as to realize not only the editing of the second target object in the virtual scene, but also the editing of the second object generation function required to construct the second target object, improving the flexibility of virtual scene editing and the efficiency of generating virtual scenes.

[0169] In some embodiments, this embodiment further includes:

[0170] Obtain object function information;

[0171] Through the large language model, based on the object function information, generate a sample function call instruction, and based on the sample function call instruction, generate an object generation function required to construct an object.

[0172] Among them, the object function information refers to the information used to indicate the generation of the object generation function. The object generation function is a function used to generate objects in the virtual scene. The object generation function includes a first object generation function and / or a second object generation function.

[0173] Optionally, various sample function call instructions of the object generation function can be generated through the large language model based on the object function information. For example, the sample function call instructions generated based on the object function information can include sample function call instructions corresponding to bungalows, villas, rockeries, fountains, trees, tables and chairs, pyramids, trees, snow mountains, farmland, and fountains. The object generation functions generated based on the sample function call instructions can be used to construct bungalows, villas, rockeries, fountains, trees, tables and chairs, pyramids, trees, snow mountains, farmland, and fountains respectively.

[0174] Optionally, the terminal can generate function prompt information based on the object function information, and input the function prompt information into the large language model to obtain the object generation function.

[0175] Optionally, after obtaining the object generation function, the user can also improve and adjust the object generation function to obtain the final object generation function.

[0176] In this embodiment, object function information is obtained. Through a large language model, based on the object function information, an example of a function call instruction is generated, and based on the example of the function call instruction, an object generation function required for constructing an object is generated, realizing the generation of an object generation function through a large language model, without the need for the user to manually generate the object generation function, and improving the efficiency of generating the object generation function.

[0177] The following is a further description of the scenario processing method provided by this application according to Figure 3 . In this embodiment, a game scenario is used as an example of a virtual scenario for illustration. The game scenario is displayed on the graphical user interface through the game client. In response to the input operation of the player on the graphical user interface, first object description information is obtained. The first object description information is used for a first target object. Through a classifier, based on the first object description information, the object type to which the first target object belongs is determined.

[0178] When the object type is the first target type, feature extraction is performed on the first description information, and based on the extracted features, the material information required for constructing the first target object is determined from the second database. When the object type is the first target type, the first function information required for constructing the first target object is determined from the first database, the first object description information is refined to obtain second object description information, example information is obtained from the first database based on the first object description information, and prompt information is generated based on the second object description information, the material information, the first function information, the information indicating the output thinking process, and the example information.

[0179] When the object type is the second target type, example information is obtained from the second database based on the first object description information, the first object description information is refined to obtain second object description information, and prompt information is generated based on the second object description information, the information indicating the output thinking process, and the example information.

[0180] Through a large language model, at least one first function call instruction required for constructing the first target object is generated based on the prompt information, and the corresponding first object generation function is called based on the at least one first function call instruction to generate the first target object in the virtual scenario.

[0181] Among them, when the object type is the second target type, it indicates that there is a first target object that has been constructed in the second database. At this time, the first target object can be retrieved from the second database. Optionally, the first target object can be retrieved from the database through a retrieval enhancement generation method. The second database can be, for example, Vector Database 2, and the first database can be, for example, Knowledge Base 1 or Vector Database 1. The same data can be stored in Knowledge Base 1 and Vector Database 1. However, the data in Knowledge Base 1 exists in scalar form, and the data in Vector Database 1 exists in vector form.

[0182] Optionally, when the object type is the second target type, since there is a first target object that has been constructed in the second database, the first object generation function can default to a loading function, and the first function call instruction can be a call instruction for the loading function. At this time, the first target object can be directly loaded into the game scene through the loading function. Therefore, when the object type is the second target type, it is not necessary to generate a prompt message based on the first function information.

[0183] In this embodiment, the first target object is generated by combining a large language model, COT reasoning (corresponding to the information indicating the thinking process), and ICL technology (corresponding to example information), realizing the automatic construction of objects by the intelligent agent, and improving the efficiency and accuracy of the player in constructing objects. Moreover, in this embodiment, a database is constructed, objects are classified, and features are extracted, making the generated first target object more accurate.

[0184] Next, according to Figure 4 , the scene processing method provided by this application will be further described.

[0185] The game scene is displayed in the graphical user interface through the game editor. In response to a second input operation on the graphical user interface, object function information is obtained. Through the large language model, based on the object function information, a sample of function call instructions is generated, and based on the sample of function call instructions, an object generation function required for constructing an object is generated. In response to a first input operation on the graphical user interface, object position adjustment information of a second target object in the virtual scene is obtained. Based on the object position adjustment information, at least one second function call instruction required for editing the second target object is generated. The corresponding second object generation function is called based on the at least one second function call instruction to adjust the position of the second target object in the game scene. In response to a first input operation on the graphical user interface, the function code corresponding to the second object generation function required for constructing the second target object in the virtual scene is obtained, and the function code is edited through the large language model to edit the second object generation function corresponding to the game scene.

[0186] Among them, after adjusting the position of the second target object, the layout mode of the second target object may or may not change. Editing the function code may involve adding function code, modifying function code, or deleting function code. When adding function code, a corresponding target object may or may not be added to the game scene.

[0187] In this embodiment, an Artificial Intelligence Generated Content (AIGC) framework is constructed to enable development staff to build objects in the game scene with the assistance of a large language model, reducing repetitive development work and improving the development efficiency of the game scene.

[0188] From the above, it can be seen that in the embodiment of this application, in response to an input operation on the graphical user interface, first object description information is obtained. The first object description information is used to describe the first target object. Based on the first object description information, a prompt message is generated, and at least one first function call instruction required to build the first target object is generated based on the prompt message through a large language model. Based on the at least one first function call instruction, the corresponding first object generation function is called to generate the first target object in the virtual scene, enabling the user to generate the first target object in the virtual scene by inputting the first object description information on the graphical user interface. Even if the user is not familiar with the construction method and / or the object to be constructed is complex, the first target object can be quickly generated, reducing the time to generate the first target object and improving the efficiency of generating the first target object.

[0189] To better implement the above method, the embodiment of this application further provides a scene processing device. The scene processing device may be specifically integrated in an electronic device, such as a computer device, which may be a terminal, a server, or other devices.

[0190] Among them, the terminal may be a mobile phone, a tablet computer, a smart Bluetooth device, a laptop computer, a personal computer, or other devices; the server may be a single server or a server cluster composed of multiple servers.

[0191] For example, in this embodiment, taking the scene processing device being specifically integrated in the terminal as an example, the method of the embodiment of this application will be described in detail. This embodiment provides a scene processing device that displays a virtual scene through a graphical user interface. The scene processing device is as Figure 5 shown. The scene processing device may include:

[0192] An information acquisition module 501, configured to obtain first object description information in response to an input operation on the graphical user interface. The first object description information is used to describe the first target object.

[0193] The instruction generation module 502 is configured to generate a prompt message based on the first object description information, and generate at least one first function call instruction required to construct the first target object based on the prompt message through a large language model.

[0194] The object generation module 503 is configured to call the corresponding first object generation function based on at least one first function call instruction to generate the first target object in the virtual scene.

[0195] In some embodiments, the instruction generation module 502 is specifically configured to:

[0196] Determine the object type to which the first target object belongs based on the first object description information;

[0197] When the object type is the first target type, obtain the material information required to construct the first target object;

[0198] Generate a prompt message based on the first object description information and the material information.

[0199] In some embodiments, the instruction generation module 502 is specifically configured to:

[0200] Determine the first function information required to construct the first target object from the first database based on the object type;

[0201] Generate a prompt message based on the first object description information, the material information, and the first function information.

[0202] In some embodiments, the instruction generation module 502 is specifically configured to:

[0203] Perform a refinement process on the first object description information to obtain the second object description information;

[0204] Generate a prompt message based on the second object description information.

[0205] In some embodiments, the instruction generation module 502 is specifically configured to:

[0206] Obtain example information from the second database;

[0207] Generate a prompt message based on the first object description information and the example information.

[0208] In some embodiments, the above scene processing device further includes an editing module, and the editing module is specifically configured to:

[0209] In response to a first input operation on the graphical user interface, obtain scene editing information;

[0210] Edit the virtual scene based on the large language model and the scene editing information.

[0211] In some embodiments, the scene editing information includes object position adjustment information of a second target object in the virtual scene, and the editing module is specifically configured to:

[0212] Through a large language model, based on the object position adjustment information, generate at least one second function call instruction required for editing the second target object;

[0213] Based on at least one second function call instruction, call the corresponding second object generation function to adjust the position of the second target object in the virtual scene.

[0214] In some embodiments, the scene editing information includes function codes corresponding to the second object generation function required to construct the second target object in the virtual scene, and the editing module is specifically configured to:

[0215] Through a large language model, edit the function codes to edit the second object generation function corresponding to the virtual scene.

[0216] In some embodiments, the editing module is further configured to:

[0217] Obtain object function information;

[0218] Through a large language model, based on the object function information, generate a function call instruction sample, and based on the function call instruction sample, generate an object generation function required for constructing an object.

[0219] In specific implementation, the above-mentioned modules can be implemented as independent entities, or can be combined arbitrarily to be implemented as the same or several entities. For the specific implementation manners of the above-mentioned modules and the corresponding beneficial effects, reference can be made to the method embodiments described above, which will not be elaborated herein.

[0220] Correspondingly, an embodiment of the present application further provides an electronic device, which can be a terminal, and the terminal can be a terminal device such as a smart phone, a tablet computer, a notebook computer, a touch screen, a game console, a personal computer (PC), a personal digital assistant (PDA), etc. As Figure 6 shown, Figure 6The figure is a schematic structural diagram of an electronic device provided by an embodiment of the present application. The electronic device 600 includes a processor 601 having one or more processing cores, a memory 602 having one or more computer-readable storage media, and a computer program stored on the memory 602 and executable on the processor. Among them, the processor 601 is electrically connected to the memory 602. Those skilled in the art can understand that the structural diagram of the electronic device shown in the figure does not constitute a limitation on the electronic device, and it may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0221] The processor 601 is the control center of the electronic device 600, connecting various parts of the entire electronic device 600 through various interfaces and lines. By running or loading software programs and / or modules stored in the memory 602, and calling data stored in the memory 602, it executes various functions of the electronic device 600 and processes data, thereby monitoring the entire electronic device 600.

[0222] In the embodiment of the present application, the processor 601 in the electronic device 600 will load the instructions corresponding to the processes of one or more application programs into the memory 602 according to the following steps, and the processor 601 will run the application programs stored in the memory 602 to implement various functions, such as:

[0223] In response to an input operation on the graphical user interface, obtain first object description information, where the first object description information is used to describe a first target object;

[0224] Based on the first object description information, generate a prompt message, and based on the prompt message through a large language model, generate at least one first function call instruction required to construct the first target object;

[0225] Based on at least one first function call instruction, call the corresponding first object generation function to generate the first target object in the virtual scene.

[0226] For the specific implementation manners of the above operations and the corresponding beneficial effects, reference can be made to the detailed description of the scene processing method above, which will not be elaborated here.

[0227] Optionally, as Figure 6 shown, the electronic device 600 further includes: a touch display screen 603, a radio frequency circuit 604, an audio circuit 605, an input unit 606, and a power supply 607. Among them, the processor 601 is electrically connected to the touch display screen 603, the radio frequency circuit 604, the audio circuit 605, the input unit 606, and the power supply 607 respectively. Those skilled in the art can understand, Figure 6The structure of the electronic device shown does not limit the electronic device, and it may include more or fewer components than shown, or combine certain components, or have a different component arrangement.

[0228] The touch display screen 603 can be used to display a graphical user interface and receive operation instructions generated by the user acting on the graphical user interface. The touch display screen 603 may include a display panel and a touch panel. Among them, the display panel can be used to display information input by the user or information provided to the user, as well as various graphical user interfaces of the electronic device. These graphical user interfaces can be composed of graphics, text, icons, videos, and any combination thereof. Optionally, the display panel can be configured in the form of a liquid crystal display (LCD), an organic light-emitting diode (OLED), etc. The touch panel can be used to collect touch operations of the user on or near it (such as operations of the user using a finger, a stylus, or any suitable object or accessory on or near the touch panel), and generate corresponding operation instructions, and the operation instructions execute the corresponding program. Optionally, the touch panel can include two parts: a touch detection device and a touch controller. Among them, the touch detection device detects the touch orientation of the user, detects the signal brought by the touch operation, and transmits the signal to the touch controller; the touch controller receives the touch information from the touch detection device, converts it into contact coordinates, and then sends it to the processor 601, and can receive and execute the commands sent by the processor 601. The touch panel can cover the display panel. When the touch panel detects a touch operation on or near it, it is transmitted to the processor 601 to determine the type of touch event. Subsequently, the processor 601 provides a corresponding visual output on the display panel according to the type of touch event. In the embodiments of the present application, the touch panel and the display panel can be integrated into the touch display screen 603 to implement input and output functions. However, in some embodiments, the touch panel and the touch panel can be implemented as two independent components to implement input and output functions. That is, the touch display screen 603 can also be used as a part of the input unit 606 to implement the input function.

[0229] The radio frequency circuit 604 can be used to transmit and receive radio frequency signals to establish wireless communication with a network device or other electronic devices through wireless communication, and transmit and receive signals with the network device or other electronic devices.

[0230] The audio circuit 605 can be used to provide an audio interface between the user and the electronic device through a speaker and a microphone. The audio circuit 605 can transmit the electrical signal converted from the received audio data to the speaker, which converts it into a sound signal for output; on the other hand, the microphone converts the collected sound signal into an electrical signal, which is received by the audio circuit 605, converted into audio data, and then the audio data is output to the processor 601 for processing, and then sent to another electronic device, for example, through the radio frequency circuit 604, or the audio data is output to the memory 602 for further processing. The audio circuit 605 may also include an earphone jack to provide communication between the peripheral earphone and the electronic device.

[0231] The input unit 606 can be used to receive input digital, character information or user characteristic information (such as fingerprint, iris, face information, etc.), and generate keyboard, mouse, joystick, optical or trackball signal inputs related to user settings and function controls.

[0232] The power supply 607 is used to supply power to each component of the electronic device 600. Optionally, the power supply 607 can be logically connected to the processor 601 through a power management system, so as to realize functions such as management of charging, discharging, and power consumption management through the power management system. The power supply 607 may also include any components such as one or more DC or AC power supplies, a recharge system, a power failure detection circuit, a power converter or inverter, and a power status indicator.

[0233] Although Figure 6 not shown in the figure, the electronic device 600 may also include a camera, a sensor, a Wi-Fi module, a Bluetooth module, etc., which will not be elaborated here.

[0234] In the above embodiments, the descriptions of the various embodiments have their own emphases. For the parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0235] Those of ordinary skill in the art can understand that all or part of the steps in the above various methods can be completed by instructions, or by controlling relevant hardware through instructions. The instructions can be stored in a computer-readable storage medium and loaded and executed by a processor.

[0236] Therefore, an embodiment of the present application provides a computer-readable storage medium, in which multiple computer programs are stored. The computer programs can be loaded by a processor to execute any one of the scenario processing methods provided by the embodiments of the present application. For example, the computer program can execute the following steps:

[0237] In response to an input operation on the graphical user interface, obtain first object description information, where the first object description information is used to describe a first target object;

[0238] Generate a prompt message based on the first object description information, and generate at least one first function call instruction required to construct the first target object based on the prompt message through a large language model;

[0239] Call the corresponding first object generation function based on at least one first function call instruction to generate the first target object in the virtual scene.

[0240] For the specific implementation manners of the above operations and the corresponding beneficial effects, reference may be made to the detailed description of the scene processing method above, which will not be elaborated here.

[0241] Among them, the computer-readable storage medium may include: read-only memory (ROM, Read Only Memory), random access memory (RAM, Random Access Memory), a magnetic disk or an optical disc, etc.

[0242] Since the computer program stored in the computer-readable storage medium can execute any one of the scene processing methods provided by the embodiments of the present application, the beneficial effects that can be achieved by any one of the scene processing methods provided by the embodiments of the present application can be realized. For details, see the previous embodiments and will not be elaborated here.

[0243] The above has introduced in detail a scene processing method, apparatus, electronic device, and computer-readable storage medium provided by the embodiments of the present application. Specific examples are used in this article to elaborate on the principle and implementation manner of the present application. The description of the above embodiments is only used to help understand the method and its core idea of the present application; at the same time, for those skilled in the art, according to the idea of the present application, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present application.

Claims

1. A scene processing method, characterized in that, Display a virtual scene through a graphical user interface, the method comprising: In response to an input operation on the graphical user interface, obtain first object description information for describing a first target object; Based on the first object description information, generate a prompt message, and based on the prompt message through a large language model, generate at least one first function call instruction required to construct the first target object; Based on the at least one first function call instruction, call a corresponding first object generation function to generate the first target object in the virtual scene.

2. The scenario processing method according to claim 1, wherein The method further comprises: Based on the first object description information, determine the object type to which the first target object belongs; In the case where the object type is a first target type, obtain material information required to construct the first target object; The generating a prompt message based on the first object description information includes: Generating a prompt message based on the first object description information and the material information.

3. The scene processing method according to claim 2, characterized in that The method further comprises: Based on the object type, determine first function information required to construct the first target object from a first database; The generating a prompt message based on the first object description information and the material information includes: Generating a prompt message based on the first object description information, the material information, and the first function information.

4. The scenario processing method according to claim 1, characterized in that The generating a prompt message based on the first object description information includes: Perform a refinement process on the first object description information to obtain second object description information; Generate a prompt message based on the second object description information.

5. The scenario processing method according to claim 1, wherein The generating a prompt message based on the first object description information includes: Obtain example information from a second database; Generate a prompt message based on the first object description information and the example information.

6. The scenario processing method according to claim 1, wherein, The method further comprises: In response to a first input operation on the graphical user interface, obtain scene editing information; Based on the large language model and the scene editing information, edit the virtual scene.

7. The scenario processing method according to claim 6, wherein The scene editing information includes object position adjustment information of a second target object in the virtual scene, and the editing the virtual scene based on the large language model and the scene editing information includes: Through the large language model, generate at least one second function call instruction required to edit the second target object based on the object position adjustment information; Based on the at least one second function call instruction, call a corresponding second object generation function to adjust the position of the second target object in the virtual scene.

8. The scene processing method according to claim 6, characterized in that The scene editing information includes function code corresponding to a second object generation function required to construct a second target object in the virtual scene, and the editing the virtual scene based on the large language model and the scene editing information includes: Through the large language model, edit the function code to edit the second object generation function corresponding to the virtual scene.

9. The scene processing method according to any one of claims 1-8, characterized in that, The method further comprises: Obtain object function information; Through the large language model, based on the object function information, generate sample function call instructions, and based on the sample function call instructions, generate an object generation function required for constructing an object.

10. A scene processing device, characterized in that, Display a virtual scene through a graphical user interface, the device comprising: An information acquisition module, configured to acquire first object description information in response to an input operation on the graphical user interface, where the first object description information is used to describe a first target object; An instruction generation module, configured to generate a prompt message based on the first object description information, and through the large language model, generate at least one first function call instruction required for constructing the first target object based on the prompt message; An object generation module, configured to call a corresponding first object generation function based on the at least one first function call instruction to generate the first target object in the virtual scene.

11. An electronic device, characterized in that, Comprising a processor and a memory, the memory stores multiple instructions; the processor loads the instructions from the memory to execute the scene processing method according to any one of claims 1 to 9.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores multiple instructions, and the instructions are suitable for being loaded by a processor to execute the scene processing method according to any one of claims 1 to 7.

Images