Data processing method and electronic equipment

Abstract

The invention discloses a data processing method and electronic equipment, and the method comprises the steps: obtaining a target processing task in response to first processing equipment, and determining the target processing equipment from an equipment cluster based on first reference information, the equipment cluster being an equipment group where the first processing equipment and the target processing equipment are located; determining a target communication path from the first processing device to the target processing device based on second reference information, the target communication path passing through at least one node device except the first processing device and the target processing device in the device cluster, the node equipment and the first processing equipment are located at the same or different network levels; and transmitting target task data of the target processing task to the target processing equipment based on the target communication path.

Description

Technical Field

[0001] This application relates to the technical field of data processing, and particularly relates to a data processing method and an electronic device. Background Art

[0002] In the computing power sharing scenario, nodes are usually selected nearby, and the data to be processed is given to the nearby nodes for data processing. However, if the current nearest node is in the state of data processing, waiting is required, and it is easy to cause traffic bandwidth blockage of the current node or the upstream node, while other nodes are in the idle state, and the computing power resources of other nodes in the network are not fully utilized, affecting the processing efficiency of the computing power cluster. Summary of the Invention

[0003] Embodiments of this application provide a data processing method and an electronic device.

[0004] On the one hand, embodiments of this application provide a data processing method, including:

[0005] In response to the first processing device obtaining a target processing task, determining a target processing device from a device cluster based on first reference information, where the device cluster is a device group where the first processing device and the target processing device are located;

[0006] Determining a target communication path from the first processing device to the target processing device based on second reference information, where the target communication path passes through at least one node device in the device cluster other than the first processing device and the target processing device, and the node device is at the same or different network levels as the first processing device;

[0007] Transmitting the target task data of the target processing task to the target processing device based on the target communication path.

[0008] Optionally, determining a target processing device from a device cluster based on first reference information includes:

[0009] Obtaining the device configuration information of each processing device in the device cluster, where the device configuration information includes at least one of hardware configuration information, software configuration information, and model configuration information;

[0010] Determining a target processing device from the device cluster by using a first processing model built in the first processing device based on the device configuration information and the task execution information of each processing device.

[0011] Optionally, determining a target processing device from the device cluster by using a first processing model built in the first processing device based on the device configuration information and the task execution information of each processing device includes:

[0012] Determine the execution weights of each processing device for the target processing task based on the device configuration information, the task execution information, and the task information of the target processing task;

[0013] Determine at least one processing device from the device cluster as the target processing device based on the execution weights.

[0014] Optionally, determining a target processing device from the device cluster based on first reference information includes at least one of the following:

[0015] Sort the processing devices in the device cluster based on the resource usage of each processing device and the positional relationship with the first processing device, and determine the second processing device in the first sequence as the target processing device;

[0016] Obtain the user profile information of the target user and the location information of the device cluster, and determine the target processing device from the device cluster based on the user profile information and the location information;

[0017] Determine the target processing device from the device cluster based on the network usage of each processing device in the device cluster and the network communication status between the processing devices;

[0018] Obtain the task information of the target processing task and the model information configured for each processing device in the device cluster, and determine the target processing device based on the task information and the model information;

[0019] Obtain the selection operation of the target user, and determine the fourth processing device in the device cluster that matches the selection operation as the target processing device.

[0020] Optionally, the determining the target communication path from the first processing device to the target processing device based on second reference information includes:

[0021] Obtain the communication link status between each node on the communication path from the first processing device to the target processing device in the device cluster;

[0022] Evaluate the communication path between the first processing device and the target processing device based on the communication link status;

[0023] Determine the target communication path based on the evaluation result.

[0024] Optionally, evaluating the communication path between the first processing device and the target processing device based on the communication link status includes:

[0025] Determine the configuration weights of the communication paths between each node based on the communication link status between the nodes;

[0026] Evaluate the data transmission cost of the communication path from the first processing device to the target processing device through at least one adjacent node device based on the configuration weights, so as to determine the target communication path based on the data transmission cost.

[0027] Optionally, determining the target communication path based on the evaluation result includes:

[0028] Evaluate the first data transmission cost of the first communication path between the first processing device and its adjacent nodes based on the configuration weights of the first communication path, so as to determine the first node device based on the first data transmission cost;

[0029] Evaluate the second data transmission cost of the second communication path between the first node device and its adjacent nodes based on the configuration weights of the second communication path, so as to determine the second node device based on the second data transmission cost;

[0030] Repeat the step of determining the target node device based on the configuration weights of the communication paths between each node device and its adjacent nodes, so as to determine the target communication path from the first processing device to the target processing device.

[0031] Optionally, determining the target communication path based on the evaluation result includes:

[0032] Evaluate the first data transmission cost of the first communication path between the first processing device and its adjacent nodes based on the configuration weights of the first communication path, so as to determine the first node device based on the first data transmission cost;

[0033] Evaluate the second data transmission cost of the second communication path between the first node device and its adjacent nodes based on the configuration weights of the second communication path, so as to evaluate the first node device based on the second data transmission cost and the first data transmission cost;

[0034] Repeat the step of evaluating the parent node device in the previous round, so as to determine the communication path with the minimum data transmission cost as the target communication path, where the parent node is the node device determined from the adjacent nodes to construct the target communication path.

[0035] Optionally, the method further includes:

[0036] Add the first node device and its adjacent node devices to the first list, and the node devices included in the first list are the node devices that have been checked;

[0037] When replacing the parent node device, determine whether the current parent node device and its adjacent node devices are in the first list;

[0038] When the current parent node device and its adjacent node devices are not in the first list, update the first list based on the current parent node device and its adjacent node devices.

[0039] On the other hand, an embodiment of the present application also provides an electronic device,

[0040] When it is used as the first processing device, it includes at least one processor and at least one processing model that can run on the processor. The processing model can be called by a target application to perform at least one of the following:

[0041] In response to the first processing device obtaining a target processing task, determine a target processing device from a device cluster based on first reference information. The device cluster is the device group where the first processing device and the target processing device are located;

[0042] Determine a target communication path from the first processing device to the target processing device based on second reference information. The target communication path passes through at least one node device in the device cluster other than the first processing device and the target processing device. The node device is at the same or different network levels as the first processing device;

[0043] Transmit the target task data of the target processing task to the target processing device based on the target communication path. Description of the Drawings

[0044] Figure 1 is a flowchart of the data processing method according to an embodiment of the present application;

[0045] Figure 2 is of an embodiment of the present application Figure 1 a flowchart of an embodiment of step S100;

[0046] Figure 3 is of an embodiment of the present application Figure 2 a flowchart of an embodiment of step S120;

[0047] Figure 4 is of an embodiment of the present application Figure 1 a flowchart of an embodiment of step S200;

[0048] Figure 5 is of an embodiment of the present application Figure 4 a flowchart of an embodiment of step S220;

[0049] Figure 6For an embodiment of the present application Figure 4 Flowchart of an embodiment of step S230 in

[0050] Figure 7 For an embodiment of the present application Figure 4 Flowchart of another embodiment of step S230 in

[0051] Figure 8 Flowchart of a specific embodiment for determining data transmission cost in an embodiment of the present application

[0052] Figure 9 Flowchart of another specific embodiment for determining data transmission cost in an embodiment of the present application

[0053] Figure 10 Flowchart of yet another specific embodiment for determining data transmission cost in an embodiment of the present application

[0054] Figure 11 Flowchart of yet another specific embodiment for determining data transmission cost in an embodiment of the present application

[0055] Figure 12 Flowchart of yet another specific embodiment for determining data transmission cost in an embodiment of the present application

[0056] Figure 13 Flowchart of yet another specific embodiment for determining data transmission cost in an embodiment of the present application

[0057] Figure 14 Flowchart of yet another specific embodiment for determining data transmission cost in an embodiment of the present application

[0058] Figure 15 Schematic structural diagram of an electronic device in an embodiment of the present application Detailed implementation manners

[0059] Various solutions and features of the present application are described herein with reference to the accompanying drawings. It should be understood that various modifications can be made to the embodiments of the application herein. Therefore, the above description should not be regarded as a limitation, but only as an example of the embodiments. Those skilled in the art will think of other modifications within the scope and spirit of the present application.

[0060] The accompanying drawings included in and constituting a part of the specification illustrate embodiments of the present application and, together with the general description of the present application given above and the detailed description of the embodiments given below, are used to explain the principles of the present application.

[0061] These and other features of the present application will become apparent from the following description of the preferred forms of the embodiments given as non - limiting examples with reference to the accompanying drawings.

[0062] It should also be understood that although the present application has been described with reference to some specific examples, those skilled in the art can surely implement many other equivalent forms of the present application.

[0063] When combined with the accompanying drawings, the above and other aspects, features, and advantages of the present application will become more apparent in view of the following detailed description.

[0064] Specific embodiments of the present application will be described hereinafter with reference to the accompanying drawings; however, it should be understood that the claimed embodiments are merely examples of the present application and can be implemented in various ways. Well-known and / or repetitive functions and structures are not described in detail to avoid obscuring the present application with unnecessary or redundant details. Therefore, the specific structural and functional details claimed herein are not intended to be limiting, but rather are merely a basis and representative basis for the claims to teach those skilled in the art to use the present application in substantially any suitable detailed structure in a variety of ways.

[0065] This specification may use the phrases "in one embodiment", "in another embodiment", "in yet another embodiment", or "in other embodiments", which may each refer to one or more of the same or different embodiments according to the present application.

[0066] Figure 1 A flowchart of a data processing method according to an embodiment of the present application is shown. A data processing method provided by an embodiment of the present application, as Figure 1 shown, includes:

[0067] S100, in response to a first processing device obtaining a target processing task, determining a target processing device from a device cluster based on first reference information, where the device cluster is a device group where the first processing device and the target processing device are located;

[0068] The data processing method of this embodiment is applied to a device cluster, which is a device group where the first processing device and the target processing device are located. The device cluster can be a computing power device cluster that provides computing power, or a network device cluster including non-AI devices, or an edge network cluster composed of non-computing power devices and ordinary devices. The first processing device is the initial node in the device cluster, that is, the device that interacts with the user or receives the processing task created by the user. In this embodiment, first, the first processing device obtains the user's target processing task. The target processing task can be an AI processing task. For example, a task of generating an image, an image processing task, a task of generating text, an intelligent question and answer search task, a task of generating a video or special effects, an AI editing task, an AI task of generating a data report or PPT. After the first processing device obtains the target processing task, it determines the target processing device from the device cluster based on the first reference information. The first reference information can be the idle state information of each node in the cluster within the target duration, or the computing power resource information, or the AI service information that can be provided; the target processing device can be a device used to execute at least part of the tasks of the target processing task. After the first processing device obtains the AI processing task, it needs other nodes in the device cluster to help provide computing power, and can determine the target processing device that is idle and can provide computing power support within the target duration from the device cluster based on the first reference information.

[0069] For example, the first processing device obtains an AI task of generating and processing an image. The computing power of the first processing device itself cannot complete this task independently and needs other nodes in the network device cluster to help provide computing power to complete the task. The first processing device can determine the target processing device that can provide computing power in the network device cluster according to the device information that each node in the network device cluster is idle and can provide computing power support within a certain period in the future, and execute at least part of the AI tasks of generating and processing the image through the target processing device.

[0070] S200, determine the target communication path from the first processing device to the target processing device based on the second reference information. The target communication path passes through at least one node device in the device cluster other than the first processing device and the target processing device, and the node device is at the same or different network levels as the first processing device;

[0071] In this embodiment, after determining the target processing device from the device cluster, the target communication path between the first processing device and the target processing device is determined based on the second reference information. The second reference information may be the status of the communication links between the nodes, such as bandwidth, occupancy rate, occupancy duration, and predicted congestion status change information; the target communication path is the fastest communication path between the first processing device and the target processing device. The target communication path passes through at least one node device in the device cluster other than the first processing device and the target processing device. The node device may have AI processing capabilities or only network data relay capabilities. The node device may be only used to transmit the target task data or result to the next node device or the target processing device. The node device and the first processing device are at the same or different network levels. The node device and the first processing device may be at the communication network level, such as a local area network or an external network, or at the hierarchical structure in the network structure formed by the device connection relationship.

[0072] In the case where the task processed by the first processing device requires the help of other nodes in the device cluster to provide computing power to complete, the fastest target communication path between the first processing device and the target processing device that provides computing power can be determined based on the second reference information. For example, when the first processing device obtains an AI task of generating and processing images and determines the target processing device that can provide computing power in the network device cluster, it can select the communication links that are not in the congested state based on the bandwidth status of the communication links between the nodes in the network device cluster, and form the fastest communication path between the first processing device and the target processing device, that is, the target communication path. The node devices on the target communication path and the first processing device may be in the same network level local area network. The target task data or result can be transmitted through the node devices on the target communication path until the target task data or result is transmitted to the target processing device.

[0073] S300, transmit the target task data of the target processing task to the target processing device based on the target communication path.

[0074] In this embodiment, after determining the fastest target communication path between the first processing device and the target processing device, the target task data of the target processing task is transmitted to the target processing device based on the target communication path, so as to use the target processing device to execute the target processing task; and / or, receive the task processing result of the target processing task transmitted by the target processing device based on the target communication path. The target task data may be all the data corresponding to the target processing task or part of the data corresponding to the target processing task; the target task data of the target processing task may be transmitted to the target processing device through the target communication path, so that the target processing device processes the target task data, and then executes the target processing task; after the target processing device processes the target processing task to obtain the task processing result, the task processing result may be received by the node device on the target communication path, and then the task processing result is transmitted to the first processing device.

[0075] For example, the first processing device obtains an AI task of generating an image and processing the image, determines a target processing device that can provide computing power in the network device cluster, and determines the fastest target communication path between the first processing device and the target processing device. The data corresponding to the AI task of generating an image and processing the image is transmitted to the target processing device through the target communication path, so that the target processing device processes the data corresponding to the AI task of generating an image and processing the image; after the target processing device processes the data corresponding to the AI task of generating an image and processing the image to obtain the task processing result, the task processing result can be received by the node device on the target communication path and transmitted to the first processing device.

[0076] Through the above method, the present application determines the target processing device from the device cluster based on the idle state information and computing power resources of each node in the device cluster in the future duration, and determines the fastest target communication path from the first processing device to the target processing device according to the state of the communication links between the nodes. Then, the target task data of the target processing task is transmitted to the target processing device for processing based on the target communication path. The present application makes full use of the computing power resources of the node devices in the idle state, avoids the congestion of the transmission of AI task information, and improves the processing efficiency of the computing power cluster.

[0077] In an embodiment of the present application, as Figure 2 shown, determining the target processing device from the device cluster based on the first reference information includes:

[0078] S110, obtaining the device configuration information of each processing device in the device cluster, where the device configuration information includes at least one of hardware configuration information, software configuration information, and model configuration information;

[0079] In this embodiment, in the process of selecting a target processing device that is idle within a target duration and can provide computing power support from a device cluster, the device configuration information of each processing device in the device cluster can be obtained, and based on the device configuration information of each processing device in the device cluster, a target processing device can be determined from the device cluster. The device configuration information includes at least one of hardware configuration information, software configuration information, and model configuration information. The hardware configuration information may include device information of the processor of the processing device, graphics card information, memory device information, and storage device information; the software configuration information may include operating system information of the processing device, function information of the processor, memory function information, storage function information, and graphics card function information; the model configuration information may include relevant configuration information of the large model installed and deployed in the processing device. For example, in the process of selecting a target processing device from the device cluster according to the idle state information and computing power information of each node in the cluster within the target duration, the device information and function information of the processors of each processing device in the device cluster, as well as the relevant configuration information of the large model, can be obtained, and a processing device with the AI computing power function for processing the target processing task can be determined.

[0080] S120. Use the first processing model built in the first processing device to determine a target processing device from the device cluster based on the device configuration information and the task execution information of each processing device.

[0081] In this embodiment, in the process of selecting a target processing device that is idle within a target duration and can provide computing power support from a device cluster, a target processing device can be determined from the device cluster by using the first processing model built in the first processing device based on the device configuration information and the task execution information of each processing device. Specifically, the first processing model built in the first processing device can predict the processing duration required for each processing device to execute the target processing task. The device configuration information of each processing device in the device cluster can be obtained through the first processing device, and then the first processing model is used to make a prediction based on the device configuration information and the task execution information of each processing device to obtain the processing duration required for each processing device to execute the target processing task, and the target processing device is determined based on the processing duration and the idle state information of each processing device. The task execution information includes the currently executed information and the to-be-executed task information. For example, the relevant configuration information of the large model of each processing device in the device cluster can be obtained through the first processing device, and the first processing model built in the first processing device is used to make a prediction based on the relevant configuration information of the large model of each processing device, the currently executed information, and the to-be-executed task information to obtain the processing duration required for each processing device to execute the target processing task. A processing device with a shorter processing duration required in the prediction result and an idle state in a future period of time is determined as the target processing device.

[0082] In one embodiment of the present application, such asFigure 3 As shown, using the first processing model built into the first processing device to determine a target processing device from the device cluster based on the device configuration information and the task execution information of each processing device includes:

[0083] S1201. Determine the execution weight of each processing device for the target processing task based on the device configuration information, the task execution information, and the task information of the target processing task;

[0084] In this embodiment, when using the first processing model to predict based on the device configuration information and the task execution information of each processing device and determining a target processing device from the device cluster, the execution weight of each processing device for the target processing task is determined based on the device configuration information, the task execution information, and the task information of the target processing task. The execution weight for the target processing task can be a feasibility score or a weight score for the processing device to execute the target processing task. Since the model parameters of the artificial intelligence processing models deployed on different processing devices, such as model size, model category, model ability, and model version, are different, the corresponding execution weights of the same processing device for different processing tasks may be different. The AI models configured in each processing device for processing tasks are different. For the same target processing task, the matching degree of different processing devices for the same task and the efficiency of processing the task are also different.

[0085] For example, after the first processing model obtains the target processing task, it uses the built-in first processing model to predict the execution weight of each processing device in the target device group for the target processing task based on the device configuration information, the task execution information, and the task information of the target processing task. Specifically, it is determined that the feasibility score or weight score of processing device A in the target device group for executing the target image generation task is 15, the feasibility score or weight score for executing the target text generation task is 28, and the feasibility score or weight score for executing the target document summarization task is 37. The appropriate processing device can be selected for task processing according to the execution weight of each processing device for the target processing task.

[0086] S1202. Determine at least one processing device from the device cluster as the target processing device based on the execution weight.

[0087] In this embodiment, after determining the execution weights of each processing device for the target processing task based on the device configuration information, task execution information, and task information of the target processing task, at least one processing device is determined from the device cluster as the target processing device based on the execution weights. Specifically, the processing devices can be sorted based on the execution weights of the processing devices for the target processing task. For example, the processing devices can be sorted in descending order of the execution weights. According to the sorting of the execution weights of each device, one or more processing devices can be selected from the device cluster as the target processing devices.

[0088] For example, in descending order of the execution weights, the processing device with the highest execution weight for the target processing task can be selected from the device cluster as the target processing device. Alternatively, multiple processing devices with relatively high execution weights can be selected from the device cluster as the target processing devices, so as to process the target processing task using multiple processing devices and improve the efficiency of the processing task. For illustration, after predicting the execution weights of each processing device in the target device group for the target processing task, it is determined that the feasibility score or weight score of processing device A in the target device group for executing the target image generation task is 15, the feasibility score or weight score for executing the target text generation task is 28, and the feasibility score or weight score for executing the target document summarization task is 37. The feasibility score or weight score of processing device B for executing the target image generation task is 45, the feasibility score or weight score for executing the target text generation task is 42, and the feasibility score or weight score for executing the target document summarization task is 23. The feasibility score or weight score of processing device C for executing the target image generation task is 40, the feasibility score or weight score for executing the target text generation task is 30, and the feasibility score or weight score for executing the target document summarization task is 30. For the target image generation task with a large amount of data, processing device B and processing device C can be selected as the target processing devices to process the target image generation task and improve the efficiency of processing the target image generation task.

[0089] In one embodiment of the present application, determining the target processing device from the device cluster based on the first reference information includes at least one of the following:

[0090] Sorting the processing devices in the device cluster based on the resource usage conditions of the processing devices and the positional relationship with the first processing device, and determining the second processing device in the first sequence as the target processing device;

[0091] In this embodiment, in the process of determining a target processing device from a device cluster based on first reference information, the resource usage of each processing device in the device cluster and the positional relationship with the first processing device can be obtained. The resource usage of a processing device can be the remaining situation of the current resources of the processing device. For example, the resource usage of a processing device can be the remaining situation of AI computing power resources; the positional relationship between each processing device and the first processing device can be a physical positional relationship or a network hierarchy relationship. By analyzing the positional relationship between each processing device and the first processing device, the communication duration of the communication link between each processing device and the first processing device can be obtained. Sort the processing devices based on the resource usage of each processing device in the device cluster and the positional relationship with the first processing device. Specifically, the sorting can be performed according to the processing capabilities and idle states of the processing devices, or the resource usage of each processing device and the positional relationship with the first processing device can be scored, and then sorted according to the scores; determine the second processing device in the first sequence as the target processing device, that is, determine the processing device ranked first in the sequence as the target processing device.

[0092] For example, after obtaining the resource usage of each processing device in the target device group and the positional relationship with the first processing device, it is determined that the score of the resource usage and the positional relationship with the first processing device of processing device A in the target device group is 19, the score of the resource usage and the positional relationship with the first processing device of processing device B is 34, and the score of the resource usage and the positional relationship with the first processing device of processing device C is 27. The processing devices in the target device group can be sorted from high to low according to the scores, and the order is processing device B, processing device C, and processing device A. Determine processing device B in the first sequence as the target processing device.

[0093] Obtain the user portrait information of the target user and the location information of the device cluster, and determine a target processing device from the device cluster based on the user portrait information and the location information;

[0094] In this embodiment, during the process of determining a target processing device from a device cluster based on first reference information, user profile information of the target user and location information of the device cluster can be obtained. The user profile information can characterize the user's historical usage data, representing the user's habits or preferences. Based on the location information in the device cluster, the type of the device cluster can be determined. Determining the type of the device cluster can ensure the privacy of the processing device. For example, the types of device clusters can include home device clusters and company device clusters, and it is necessary to ensure the privacy of the processing devices in the home device clusters and company device clusters. The target processing device is determined from the device cluster based on the user profile information and the location information of the device cluster. For example, the historical usage data of the user for generating images is determined from the user profile information, and then it is determined that the user's habit for generating images is to use a fixed model to generate high-resolution images. After obtaining the location information of the target device cluster and determining that the target device cluster is a company device cluster, a processing device with an AI model having the function of generating high-resolution images can be selected from the company device cluster as the target processing device.

[0095] The target processing device is determined from the device cluster based on the network usage conditions of each processing device in the device cluster and the network communication status among the processing devices.

[0096] In this embodiment, during the process of determining a target processing device from a device cluster based on first reference information, the network usage conditions of each processing device in the device cluster and the network communication status among the processing devices can be obtained. The network usage conditions of each processing device can include the network bandwidth, quality, and stability of the processing device. By obtaining the network communication status among the processing devices, the network congestion status among the node devices in the next period of time can be determined. The target processing device is determined from the device cluster based on the network usage conditions of each processing device in the device cluster and the network communication status among the processing devices. For example, after obtaining the network bandwidth of each processing device in the target device cluster and the network communication transmission rate among the processing devices, one or more processing devices with good network bandwidth quality and fast communication transmission rate can be determined as the target processing device. The first processing device can send the AI task to the target processing device at a faster transmission rate, and the target processing device with better network bandwidth quality can process the AI task.

[0097] The task information of the target processing task and the model information configured for each processing device in the device cluster are obtained, and the target processing device is determined based on the task information and the model information.

[0098] In this embodiment, in the process of determining a target processing device from a device cluster based on first reference information, task information of a target processing task and model information configured for each processing device in the device cluster can be obtained, and the target processing device is determined based on the task information and the model information. The task information of the target processing task may include task content, task type, task requirements, and an estimated task duration. The model information may include specific configurations of an AI processing model. The target processing task may be an image generation task, a text summarization processing task, or a video generation task. The model configurations and processing durations required for these different types of tasks are different, and a model adapted to the task information can be selected based on the task information of different types of tasks to process the task. For example, after obtaining the task information of the target processing task, if it is determined that the type of the target processing task is an image generation task, a processing device with an AI image generation function configuration can be selected from the device cluster according to the model information configured for each processing device in the device cluster, and the processing device with a shorter processing duration among the processing devices with an AI image generation function is determined as the target processing device.

[0099] Obtain a selection operation of a target user, and determine the fourth processing device in the device cluster that matches the selection operation as the target processing device.

[0100] In this embodiment, in the process of determining a target processing device from a device cluster based on first reference information, a selection operation of a target user can be obtained through a first processing device. The selection operation of the target user may include information corresponding to an operation performed by the target user through an interaction interface of the first processing device. For example, a selection of the type of the target processing task. After obtaining the selection operation of the target user, the fourth processing device in the device cluster that matches the selection operation is determined as the target processing device. Specifically, relevant configuration information of a large model that matches the information corresponding to the selection operation can be determined according to the information corresponding to the selection operation, and then the processing device corresponding to the relevant configuration information of the large model is determined as the target processing device. For example, if the user manually selects through the interaction interface of the first processing device that the type of the target processing task is an image generation task, relevant configuration information of a large model that matches the information type corresponding to the image generation task selected by the user can be determined, and then the processing device corresponding to the relevant configuration information of the large model for image generation is determined as the target processing device.

[0101] In an embodiment of the present application, as Figure 4 shown, determining a target communication path from the first processing device to the target processing device based on second reference information includes:

[0102] S210, obtain the communication link status between each node on the communication path from the first processing device to the target processing device in the device cluster;

[0103] In this embodiment, in the process of determining the target communication path from the first processing device to the target processing device based on the second reference information, first, the communication link status between each node on the communication path from the first processing device to the target processing device in the device cluster is obtained. The communication path from the first processing device to the target processing device can be all the communication paths between the first processing device and the target processing device, or a specified communication path or a communication path within a specified area between the first processing device and the target processing device; the communication link status between each node on the communication path from the first processing device to the target processing device can be the bandwidth, occupancy rate, occupancy duration, and predicted congestion status change information of the communication link between each node. For example, the processing devices included in the target area all have large models with fixed functions, and the communication network level between the processing devices included in the target area is the local area network level. It is possible to determine each communication path from the first processing device to the target processing device within the target area; based on the bandwidth status of the communication paths between each node in the target area, select the communication paths that are not in a congested state to ensure the smooth communication transmission between the first processing device and the target processing device.

[0104] S220, evaluate the communication path between the first processing device and the target processing device based on the communication link status;

[0105] In this embodiment, after obtaining the communication link status between each node on the communication path from the first processing device to the target processing device in the device cluster, the communication path between the first processing device and the target processing device is evaluated based on the communication link status. Specifically, the data transmission cost of the communication path between the first processing device and the target processing device is evaluated to obtain the evaluation results of each communication path between the first processing device and the target processing device for the data transmission cost, and the communication path with the minimum data transmission cost is determined according to the evaluation results. The data transmission cost can include the energy consumption of data transmission and the distance between nodes.

[0106] For example, denote the data transmission cost of the communication path between the first processing device and the target processing device as G + H = F, where H represents the distance from the current node to the next node, G represents the energy consumption cost of transmitting data from the current node to the next node, and F represents the total data transmission cost from the current node to the next node. After determining each communication path from the first processing device to the target processing device within the target area, starting from the first processing device, the transmission costs of the communication paths formed by each node between the first processing device and the target processing device can be calculated sequentially. Finally, the value of the data transmission cost F of the communication path between the first processing device and the target processing device is obtained. There can be multiple communication paths between the first processing device and the target processing device. Evaluate the values of the data transmission costs F corresponding to the multiple communication paths to determine the energy consumption and distance of data transmission for different communication paths between the first processing device and the target processing device.

[0107] S230. Determine the target communication path based on the evaluation result.

[0108] In this embodiment, after evaluating the communication paths between the first processing device and the target processing device based on the communication link state and obtaining the evaluation results of the data transmission costs for each communication path between the first processing device and the target processing device, determine the target communication path based on the evaluation results, that is, the communication path with the minimum data transmission cost. Specifically, the energy consumption and distance of data transmission among the multiple communication paths between the first processing device and the target processing device can be compared to determine the communication path with the minimum energy consumption and distance of data transmission, and this communication path is determined as the target communication path with the minimum data transmission cost. For example, after determining each communication path from the first processing device to the target processing device within the target area, starting from the first processing device, based on the transmission energy consumption cost G and distance F corresponding to the communication paths formed by each node between the first processing device and the target processing device, obtain the values of the data transmission cost F of the multiple communication paths between the first processing device and the target processing device. Among them, G + H = F. After evaluating the values of the data transmission costs F corresponding to the multiple communication paths, determine the communication path with the minimum F value among the multiple communication paths between the first processing device and the target processing device as the target communication path.

[0109] In one embodiment of the present application, as Figure 5 shown, evaluating the communication paths between the first processing device and the target processing device based on the communication link state includes:

[0110] S2201. Determine the configuration weights of the communication paths between each node based on the communication link state between each node.

[0111] In this embodiment, in the process of evaluating the communication path between the first processing device and the target processing device based on the communication link state, first, the communication link state of each node between the first processing device and the target processing device is determined. The communication link state between each node may include changes in bandwidth, occupancy rate, and congestion state. Based on the communication link state between each node, the configuration weight of the communication path between each node is determined. The configuration weight of the communication path between each node can represent the transmission energy consumption cost and distance corresponding to the communication path between each node. Based on the configuration weight of the communication path between each node, the data transmission cost of the communication path between each node can be determined.

[0112] For example, as Figures 8 - 14 shown, the second device in the second column in the target device cluster is the first processing device, and the fourth device in the fifth column is the target processing device. The configuration weight of the communication path between each node in the target device cluster can be represented by the depth of the color gray scale of the path. The communication path corresponding to the color with the maximum gray scale is the low-cost path, and the value of the transmission energy consumption cost G is 1; the communication path corresponding to the color with the minimum gray scale is the high-cost path, and the value of the transmission energy consumption cost G is 11; the communication path corresponding to the intermediate gray scale color is the obstacle path, and the value of the transmission energy consumption cost G is infinity ∞, that is, data cannot be transmitted through the obstacle path. Starting from the first processing device, the transmission energy consumption cost G and the distance F corresponding to the communication path composed of each node between the first processing device and the target processing device can be obtained, and the configuration weight of the communication path between each node is determined.

[0113] S2202, evaluate the data transmission cost of the communication path from the first processing device through at least one adjacent node device to the target processing device based on the configuration weight, so as to determine the target communication path based on the data transmission cost.

[0114] In this embodiment, after determining the configuration weight of the communication path between each node based on the communication link state between each node, evaluate the data transmission cost of the communication path from the first processing device through at least one adjacent node device to the target processing device based on the configuration weight of the communication path between each node, so as to determine the target communication path based on the data transmission cost. Specifically, obtain the transmission energy consumption cost and distance corresponding to the communication path between each node, and determine the data transmission cost of the communication path between each node according to the transmission energy consumption cost and distance corresponding to the communication path, and then determine the communication path with the minimum data transmission cost between the first processing device and the target processing device as the target communication path. For example, as Figures 8 - 13As shown in the figure, the second device in the second column of the target device cluster is the first processing device, and the fourth device in the fifth column is the target processing device. Starting from the first processing device, the values of the transmission energy consumption cost G and the distance F corresponding to the communication paths formed by each node between the first processing device and the target processing device are determined in sequence. Based on the values of the transmission energy consumption cost G and the distance F, the data transmission cost G + H = F of the communication paths formed by each node is obtained. Furthermore, the minimum data transmission cost of the communication path from the first processing device to the target processing device through at least one adjacent node device is determined, and the communication path corresponding to the minimum data transmission cost is determined as the target communication path.

[0115] In one embodiment of the present application, as Figure 6 shown, determining the target communication path based on the evaluation result includes:

[0116] S2301, evaluating the first data transmission cost of the first communication path based on the configuration weight of the first communication path between the first processing device and its adjacent nodes, so as to determine the first node device based on the first data transmission cost;

[0117] In this embodiment, after obtaining the evaluation results of the data transmission costs of each communication path between the first processing device and the target processing device, the target communication path is determined based on the evaluation results. First, the first data transmission cost of the first communication path is evaluated based on the configuration weight of the first communication path between the first processing device and its adjacent nodes. The first data transmission cost may include the transmission energy consumption cost and the distance of the first communication path between the first processing device and its adjacent nodes. Specifically, the first processing device is used as the parent node, and the adjacent nodes of the first processing device are used as the child nodes. The first data transmission cost of the first communication path between the first processing device and all adjacent nodes is determined, and the first communication path corresponding to the minimum first data transmission cost is determined. The adjacent node on this first communication path is determined as the first node device, and the data transmission cost between the first processing device and the first node device is the lowest in the first communication path between the first processing device and the adjacent nodes.

[0118] For example, as Figures 8 - 14 shown, the second device in the second column of the target device cluster is the first processing device, and the fourth device in the fifth column is the target processing device. First, the first processing device is used as the parent node, and the adjacent nodes of the first processing device are used as the child nodes. The values of the transmission energy consumption cost G and the distance F corresponding to the first communication paths between the first processing device and 4 adjacent nodes are determined in sequence. Furthermore, the first data transmission cost G + H = F of the first communication paths between the first processing device and 4 adjacent nodes is calculated, and the adjacent node corresponding to the minimum value of the first data transmission cost F can be determined as the first node device.

[0119] S2302. Evaluate the second data transmission cost of the second communication path based on the configuration weight of the second communication path between the first node device and its adjacent nodes, so as to determine a second node device based on the second data transmission cost.

[0120] In this embodiment, after evaluating the first data transmission cost of the first communication path based on the configuration weight of the first communication path between the first processing device and its adjacent nodes, and determining the first node device based on the first data transmission cost, the first node device is used as the parent node, the adjacent nodes of the first node device are used as the child nodes, the second data transmission cost of the second communication path between the first node device and all adjacent nodes is determined, and the second communication path corresponding to the minimum second data transmission cost is determined. The adjacent node on this second communication path is determined as the second node device. The data transmission cost between the first node device and the second node device is the lowest among the second communication paths between the first node device and the adjacent nodes.

[0121] For example, as Figures 8 - 14 shown, the second device in the second column in the target device cluster is the first processing device, and the fourth device in the fifth column is the target processing device. The adjacent node below the first processing device is used as the parent node, that is, the first node device, and the adjacent nodes of the first node device are used as the child nodes. The values of the transmission energy consumption cost G and the distance F corresponding to the second communication paths between the first node device and 4 adjacent nodes are determined in turn. Furthermore, the second data transmission cost G + H = F of the second communication paths between the first node device and 4 adjacent nodes is calculated. The adjacent node corresponding to the minimum value of the second data transmission cost F can be determined as the second node device.

[0122] S2303. Repeat the step of determining the target node device based on the configuration weight of the communication path between each node device and its adjacent nodes, so as to determine the target communication path from the first processing device to the target processing device.

[0123] In this embodiment, after determining the current node device on the target communication path, the step of determining the target node device based on the configuration weights of the communication paths between each node device and its adjacent nodes is repeatedly executed until the node device adjacent to the target processing device on the target communication path is determined. Specifically, the current node device is used as the parent node, the adjacent nodes of the current node device are used as the child nodes, the data transmission cost of the communication paths between the current node device and all adjacent nodes is determined, and the communication path corresponding to the minimum data transmission cost is determined. The adjacent node on this communication path is determined as the next node device, and the data transmission cost between the current node device and the next node device is the lowest among the communication paths between the current node device and the adjacent nodes. During the process of determining the node devices on the target communication path, only the shortest path nodes are determined from the adjacent nodes of the parent node, and the target path nodes in the adjacent nodes of each parent node are determined in sequence. The target communication path is composed of the communication paths between the respective target path nodes. During the process of determining the target path nodes, the relative position relationship between each node and the target processing device needs to be considered. The relative position relationship between each node and the target processing device may include the communication distance and the number of communication levels between each node and the target processing device, so as to ensure that the communication path between the target path node and the target processing device is the shortest communication path.

[0124] For example, as Figures 8 - 14 shown, the second device in the second column in the target device cluster is the first processing device, and the fourth device in the fifth column is the target processing device. Taking the first processing device as the parent node and the adjacent nodes of the first processing device as the child nodes, the first node device is determined; and after taking the first node device as the parent node and the adjacent nodes of the first node device as the child nodes to determine the second node device, the values of the transmission energy consumption cost G and the distance F corresponding to the communication paths between the current node device and 4 adjacent nodes are determined in sequence. Furthermore, the data transmission cost G + H = F of the communication paths between the current node device and 4 adjacent nodes is calculated, and the adjacent node corresponding to the minimum value of the data transmission cost F is determined as the next node device until the node device adjacent to the target processing device on the target communication path is determined. The node devices on the communication paths corresponding to each minimum transmission cost are determined as the target path nodes, and finally the communication paths between the respective target path nodes are composed into the target communication path.

[0125] In an embodiment of the present application, as Figure 7 shown, determining the target communication path based on the evaluation result includes:

[0126] S2304. Evaluate the first data transmission cost of the first communication path based on the configuration weight of the first communication path between the first processing device and its adjacent nodes, and determine the first node device based on the first data transmission cost.

[0127] In this embodiment, after obtaining the evaluation results of the data transmission costs of each communication path between the first processing device and the target processing device, the target communication path is determined based on the evaluation results. First, the first data transmission cost of the first communication path is evaluated based on the configuration weight of the first communication path between the first processing device and its adjacent nodes. The first data transmission cost may include the transmission energy consumption cost and the distance of the first communication path between the first processing device and its adjacent nodes. Specifically, taking the first processing device as the parent node and the adjacent nodes of the first processing device as the child nodes, determine the first data transmission cost of the first communication path between the first processing device and all adjacent nodes, and determine the first communication path corresponding to the minimum first data transmission cost. The adjacent node on this first communication path is determined as the first node device. The data transmission cost between the first processing device and the first node device is the lowest among the first communication paths between the first processing device and the adjacent nodes.

[0128] For example, as Figures 8 - 14 shown, the second device in the second column in the target device cluster is the first processing device, and the fourth device in the fifth column is the target processing device. First, take the first processing device as the parent node and the adjacent nodes of the first processing device as the child nodes, and sequentially determine the values of the transmission energy consumption cost G and the distance F corresponding to the first communication path between the first processing device and 4 adjacent nodes. Then calculate the first data transmission cost G + H = F of the first communication path between the first processing device and 4 adjacent nodes. The adjacent node corresponding to the minimum value of the first data transmission cost F can be determined as the first node device.

[0129] S2305. Evaluate the second data transmission cost of the second communication path based on the configuration weight of the second communication path between the first node device and its adjacent nodes, and evaluate the first node device based on the second data transmission cost and the first data transmission cost.

[0130] In this embodiment, the first data transmission cost of the first communication path is evaluated based on the configuration weight of the first communication path between the first processing device and its adjacent nodes. After determining the first node device based on the first data transmission cost, the first node device is used as the parent node, and the adjacent nodes of the first node device are used as the child nodes. The second data transmission cost of the second communication path between the first node device and all adjacent nodes is determined, and the first node device is evaluated based on the second data transmission cost and the first data transmission cost. Then, the current first node device is verified to determine whether the current first node device is the correct parent node. If the sum of the minimum transmission cost in the second data transmission cost and the first data transmission cost is the minimum data transmission cost corresponding to the communication path between the currently obtained path nodes, it can be determined that the current first node device is the correct parent node.

[0131] For example, as Figures 8 - 14 shown, the second device in the second column in the target device cluster is the first processing device, and the fourth device in the fifth column is the target processing device. The nodes adjacent to the bottom and right of the first processing device are used as the parent nodes, that is, the first node devices. The adjacent nodes of these two first node devices are used as the child nodes. The values of the transmission energy consumption cost G and the distance F corresponding to the second communication path between these two first node devices and 4 adjacent nodes are determined in sequence. Then, the second data transmission cost G + H = F of the second communication path between these two first node devices and 4 adjacent nodes is calculated. The sum of the minimum value of the calculated second data transmission cost F and the value of the first data transmission cost is 10. It can be determined that these two first node devices are the correct parent nodes.

[0132] S2306. Repeat the step of evaluating the parent node device in the previous round to determine the communication path with the minimum data transmission cost as the target communication path. The parent node is the node device that constructs the target communication path determined from the adjacent nodes.

[0133] In this embodiment, after determining the current node device on the target communication path, the steps of evaluating the parent node device in the previous round are repeatedly executed to determine the communication path with the minimum data transmission cost as the target communication path. The parent node is the node device that determines the target communication path from adjacent nodes. Specifically, the current node device is used as the parent node, and the adjacent nodes of the current node device are used as the child nodes. The minimum data transmission cost of the communication path between the current node device and all adjacent nodes is determined, and the current node device is evaluated based on this minimum data transmission cost and the minimum data transmission cost of the previous round. The current node device is verified to determine whether it is the correct parent node. If the sum of this minimum data transmission cost and the minimum data transmission cost of the previous round is the minimum data transmission cost corresponding to the communication path between the currently obtained path nodes, it can be determined that the current node device is the correct parent node.

[0134] For example, as Figures 8 - 14 shown, the fourth device in the second column and the second device in the fourth column in the target device cluster are the current parent nodes. The adjacent nodes of these two node devices are used as the child nodes. The values of the transmission energy consumption cost G and the distance F corresponding to the communication paths between these two node devices and the 4 adjacent nodes are determined in sequence. Furthermore, the data transmission cost G + H = F of the communication paths between these two node devices and the 4 adjacent nodes is calculated. The sum of the minimum value of the calculated data transmission cost F value and the minimum data transmission cost value of the previous round are 10 and 12 respectively. It can be determined that the node device on the communication path corresponding to the smaller value of these two values is the correct parent node. The next target path nodes are determined in sequence according to the correct parent node, and finally the communication paths between the respective target path nodes are formed into the target communication path.

[0135] In an embodiment of the present application, the method further includes:

[0136] Adding the first node device and its adjacent node devices to a first list, and the node devices included in the first list are the node devices that have been checked;

[0137] In the case of replacing the parent node device, determining whether the current parent node device and its adjacent node devices are in the first list;

[0138] In the case where the current parent node device and its adjacent node devices are not in the first list, updating the first list based on the current parent node device and its adjacent node devices.

[0139] In this embodiment, during the process of verifying each node device, when the parent node device is replaced, it is determined whether the current parent node device and its adjacent node devices are in the first list. The first list records the node devices that have been verified. Specifically, the first list records the data transmission cost between the verified node devices and the previous parent node device. After determining the previous parent node device and when the parent node device is replaced, it is determined whether the current parent node device and its adjacent node devices are in the first list. Specifically, it is determined whether the data transmission cost between the replaced parent node device and its adjacent node devices and the previous parent node device is recorded in the first list. If the data transmission cost between the replaced parent node device and its adjacent node devices and the previous parent node device is recorded in the first list, the data transmission cost between the replaced parent node device and its adjacent node devices and the previous parent node device can be directly obtained from the first list. If the current parent node device and its adjacent node devices are not in the first list, the data transmission cost between the current parent node device and each adjacent node device can be obtained and recorded in the first list. When determining the data transmission cost between the current parent node device and each adjacent node device, the relative position relationship between each adjacent node device and the target processing device can be determined. If the distance between a certain adjacent node device and the target processing device is relatively far, it indicates that this node device is not suitable for data transmission of the first processing device, and the data transmission cost between this adjacent node device and the current parent node device can be ignored.

[0140] For example, as Figures 8 - 14 shown, the fifth device in the second column of the target device cluster is the current parent node. Taking the adjacent nodes of this node device as child nodes, the values of the transmission energy consumption cost G and the distance F corresponding to the communication paths between this node device and 4 adjacent nodes are determined in sequence. Then, the data transmission cost G + H = F of the communication paths between this node device and 4 adjacent nodes is calculated. The node devices on the left and below this node device are not added to the first list, and the values of the data transmission cost between these two adjacent node devices and the current parent node can be added to the first list. However, the values of F of the data transmission cost between these two adjacent node devices and the current parent node are relatively high, that is, the distances between these two adjacent node devices and the target processing device are relatively far, and the data transmission cost between these two adjacent node devices and the current parent node device can be ignored.

[0141] Based on the same inventive concept, an embodiment of the present application further provides an electronic device. When it is used as the first processing device, it includes at least one processor and at least one processing model that can run on the processor. The processing model can be called by a target application to execute at least one of the following:

[0142] In response to a first processing device obtaining a target processing task, a target processing device is determined from a device cluster based on first reference information, where the device cluster is a device group where the first processing device and the target processing device are located;

[0143] A target communication path between the first processing device and the target processing device is determined based on second reference information, where the target communication path passes through at least one node device in the device cluster other than the first processing device and the target processing device, and the node device is at the same or different network levels as the first processing device;

[0144] The target task data of the target processing task is transmitted to the target processing device based on the target communication path.

[0145] Based on the same inventive concept, an embodiment of the present application further provides an electronic device, as Figure 15 shown, including:

[0146] An acquisition module configured to, in response to a first processing device obtaining a target processing task, determine a target processing device from a device cluster based on first reference information, where the device cluster is a device group where the first processing device and the target processing device are located;

[0147] A determination module configured to determine a target communication path between the first processing device and the target processing device based on second reference information, where the target communication path passes through at least one node device in the device cluster other than the first processing device and the target processing device, and the node device is at the same or different network levels as the first processing device;

[0148] A processing module configured to transmit the target task data of the target processing task to the target processing device based on the target communication path.

[0149] In an embodiment of the present application, the acquisition module is further configured to:

[0150] Obtain device configuration information of each processing device in the device cluster, where the device configuration information includes at least one of hardware configuration information, software configuration information, and model configuration information;

[0151] Determine a target processing device from the device cluster based on the device configuration information and the task execution information of each processing device using a first processing model built in the first processing device.

[0152] In an embodiment of the present application, the determination module is further configured to:

[0153] Determine the execution weights of each processing device for the target processing task based on the device configuration information, the task execution information, and the task information of the target processing task;

[0154] Determine at least one processing device from the device cluster as the target processing device based on the execution weights.

[0155] In an embodiment of the present application, the determining module is further configured to:

[0156] Sort the processing devices in the device cluster based on the resource usage of each processing device and the positional relationship with the first processing device, and determine the second processing device in the first sequence as the target processing device;

[0157] Obtain the user profile information of the target user and the location information of the device cluster, and determine the target processing device from the device cluster based on the user profile information and the location information;

[0158] Determine the target processing device from the device cluster based on the network usage of each processing device in the device cluster and the network communication status between the processing devices;

[0159] Obtain the task information of the target processing task and the model information configured for each processing device in the device cluster, and determine the target processing device based on the task information and the model information;

[0160] Obtain the selection operation of the target user, and determine the fourth processing device in the device cluster that matches the selection operation as the target processing device.

[0161] In an embodiment of the present application, the determining module is further configured to:

[0162] Obtain the communication link status between each node on the communication path from the first processing device to the target processing device in the device cluster;

[0163] Evaluate the communication path between the first processing device and the target processing device based on the communication link status;

[0164] Determine the target communication path based on the evaluation result.

[0165] In an embodiment of the present application, the determining module is further configured to:

[0166] Determine the configuration weight of the communication path between each node based on the communication link status between each node;

[0167] Evaluate the data transmission cost of the communication path from the first processing device to the target processing device through at least one adjacent node device based on the configured weights, and determine the target communication path based on the data transmission cost.

[0168] In one embodiment of the present application, the determining module is further configured to:

[0169] Evaluate the first data transmission cost of the first communication path between the first processing device and its adjacent nodes based on the configured weights of the first communication path, and determine the first node device based on the first data transmission cost;

[0170] Evaluate the second data transmission cost of the second communication path between the first node device and its adjacent nodes based on the configured weights of the second communication path, and determine the second node device based on the second data transmission cost;

[0171] Repeat the step of determining the target node device based on the configured weights of the communication paths between each node device and its adjacent nodes to determine the target communication path from the first processing device to the target processing device.

[0172] In one embodiment of the present application, the determining module is further configured to:

[0173] Evaluate the first data transmission cost of the first communication path between the first processing device and its adjacent nodes based on the configured weights of the first communication path, and determine the first node device based on the first data transmission cost;

[0174] Evaluate the second data transmission cost of the second communication path between the first node device and its adjacent nodes based on the configured weights of the second communication path, and evaluate the first node device based on the second data transmission cost and the first data transmission cost;

[0175] Repeat the step of evaluating the parent node device of the previous round to determine the communication path with the minimum data transmission cost as the target communication path, where the parent node is the node device determined from the adjacent nodes to construct the target communication path.

[0176] In one embodiment of the present application, the processing module is further configured to:

[0177] Add the first node device and its adjacent node devices to the first list, and the node devices included in the first list are the node devices that have been checked;

[0178] In the case of replacing the parent node device, determine whether the current parent node device and its adjacent node devices are in the first list;

[0179] In the case that the current parent node device and its adjacent node devices are not in the first list, update the first list based on the current parent node device and its adjacent node devices.

[0180] The above embodiments are only exemplary embodiments of the present application and are not used to limit the present application. The protection scope of the present application is defined by the claims. Those skilled in the art can make various modifications or equivalent replacements within the essence and protection scope of the present application, and such modifications or equivalent replacements should also be regarded as falling within the protection scope of the present application.

Claims

1. A data processing method, comprising: In response to the first processing device obtaining the target processing task, determining a target processing device from a device cluster based on first reference information, the device cluster being a device group where the first processing device and the target processing device are located; Determine a target communication path from the first processing device to the target processing device based on the second reference information, wherein the target communication path passes through at least one node device in the device cluster except the first processing device and the target processing device, and the node device and the first processing device are at the same or different network level; The target task data of the target processing task is transmitted to the target processing device based on the target communication path.

2. The method according to claim 1, determining a target processing device from a device cluster based on the first reference information, comprising: Obtaining device configuration information of each processing device in the device cluster, wherein the device configuration information includes at least one of hardware configuration information, software configuration information, and model configuration information; A target processing device is determined from the device cluster using a first processing model built into the first processing device based on the device configuration information and the task execution information of each processing device.

3. The method according to claim 2, wherein: Determining a target processing device from the device cluster based on the device configuration information and the task execution information of each processing device using the first processing model built in the first processing device includes: Determine the execution weight of each processing device for the target processing task based on the device configuration information, the task execution information and the task information of the target processing task; At least one processing device is determined from the device cluster as the target processing device based on the execution weight.

4. The method according to claim 1, determining the target processing device from the device cluster based on the first reference information, comprises at least one of the following: sorting the processing devices in the device cluster based on resource usage of the processing devices and their positional relationship with the first processing device, and determining a second processing device in the first sequence as the target processing device; Obtaining user portrait information of a target user and location information of the device cluster, and determining a target processing device from the device cluster based on the user portrait information and the location information; Determining a target processing device from the device cluster based on network usage of each processing device in the device cluster and network communication status between each processing device; Obtaining task information of the target processing task and model information configured for each processing device in the device cluster, and determining the target processing device based on the task information and the model information; A selection operation of a target user is obtained, and a fourth processing device in the device cluster that matches the selection operation is determined as the target processing device.

5. The method according to claim 1, wherein determining the target communication path from the first processing device to the target processing device based on the second reference information comprises: Obtaining the communication link status between each node on the communication path from the first processing device to the target processing device in the device cluster; evaluating a communication path between the first processing device and the target processing device based on the communication link status; The target communication path is determined based on the evaluation result.

6. The method according to claim 5, wherein: Evaluating a communication path between the first processing device and the target processing device based on the communication link state includes: Determine the configuration weight of the communication path between each node based on the communication link status between each node; A data transmission cost of a communication path from the first processing device through at least one adjacent node device to the target processing device is evaluated based on the configuration weight to determine the target communication path based on the data transmission cost.

7. The method according to claim 5 or 6, wherein: Determining the target communication path based on the evaluation result includes: evaluating a first data transmission cost of the first communication path based on a configuration weight of a first communication path between the first processing device and its adjacent nodes, so as to determine a first node device based on the first data transmission cost; evaluating a second data transmission cost of the second communication path based on a configuration weight of a second communication path between the first node device and its adjacent nodes, so as to determine a second node device based on the second data transmission cost; The step of determining a target node device based on a configuration weight of a communication path between each node device and its adjacent nodes is repeatedly performed to determine a target communication path from the first processing device to the target processing device.

8. The method according to claim 5 or 6, wherein: Determining the target communication path based on the evaluation result includes: evaluating a first data transmission cost of the first communication path based on a configuration weight of a first communication path between the first processing device and its adjacent nodes, so as to determine a first node device based on the first data transmission cost; evaluating a second data transmission cost of the second communication path based on a configuration weight of a second communication path between the first node device and its neighboring nodes, so as to evaluate the first node device based on the second data transmission cost and the first data transmission cost; The step of evaluating the parent node device of the previous round is repeatedly performed to determine the communication path with the minimum data transmission cost as the target communication path, wherein the parent node is a node device determined from adjacent nodes to construct the target communication path.

9. The method according to claim 5 or 6, wherein: The method further comprises: Adding the first node device and its adjacent node devices to a first list, wherein the node devices included in the first list are checked node devices; In the case of replacing the parent node device, determining whether the current parent node device and its adjacent node devices are in the first list; When the current parent node device and its neighboring node devices are not in the first list, the first list is updated based on the current parent node device and its neighboring node devices.

10. An electronic device, which, as a first processing device, comprises at least one processor and at least one processing model capable of running on the processor, wherein the processing model can be called by a target application to perform at least one of the following: In response to the first processing device obtaining the target processing task, determining a target processing device from a device cluster based on first reference information, the device cluster being a device group where the first processing device and the target processing device are located; Determine a target communication path from the first processing device to the target processing device based on the second reference information, wherein the target communication path passes through at least one node device in the device cluster except the first processing device and the target processing device, and the node device and the first processing device are at the same or different network level; The target task data of the target processing task is transmitted to the target processing device based on the target communication path.

Images