Vehicle diagnosis method of intelligent system architecture and related device

Abstract

The invention provides a vehicle diagnosis method of an intelligent system architecture and a related device. The method comprises the steps of obtaining a target diagnosis demand and historical diagnosis data of a target vehicle; labeling the historical diagnosis data according to a preset dictionary format to obtain reference data; dividing the reference data into a training set and a verification set according to a preset proportion; obtaining a pre-training model and model feature information of the pre-training model; adjusting the pre-training model according to the target diagnosis demand, the model feature information, the training set and the verification set to obtain a target model; performing system construction according to the target model to obtain a target diagnosis system; and diagnosing the target vehicle according to the target diagnosis system to obtain a target diagnosis result. The intelligent and precise vehicle diagnosis can be realized, the diagnosis cost is reduced, and the diagnosis efficiency is improved.

Description

Technical Field

[0001] This application relates to the technical field of vehicle diagnosis, and in particular, to a vehicle diagnosis method and related device with an intelligent system architecture. Background Art

[0002] With the development of automotive technology, especially the rise of new energy vehicles and intelligent networked vehicles, the field of vehicle diagnosis faces many challenges. Due to the rapid technological updates, it is necessary to continuously update equipment and diagnosis technologies to fit the characteristics of new automotive technologies. At the same time, the maintenance data and diagnostic codes of automobile manufacturers are mostly proprietary, and it is difficult for independent repair shops to obtain the latest maintenance information. Moreover, although vehicle fault phenomena are similar, the causes are diverse, which will greatly affect the accuracy and efficiency of diagnosis.

[0003] Therefore, how to improve the accuracy and efficiency of vehicle diagnosis urgently needs to be solved. Summary of the Invention

[0004] Embodiments of this application provide a vehicle diagnosis method and related device with an intelligent system architecture. By making targeted adjustments to a pre-trained model and constructing a target diagnosis system, it is possible to accurately and efficiently diagnose a vehicle, obtain reliable diagnosis results, realize the intelligence and precision of vehicle diagnosis, reduce the diagnosis cost, and improve the diagnosis efficiency.

[0005] In a first aspect, embodiments of this application provide a vehicle diagnosis method with an intelligent system architecture, and the method includes:

[0006] Obtain the target diagnosis requirements and historical diagnosis data of the target vehicle;

[0007] Label the historical diagnosis data according to a preset dictionary format to obtain reference data;

[0008] Divide the reference data into a training set and a validation set according to a preset ratio;

[0009] Obtain a pre-trained model and the model feature information of the pre-trained model;

[0010] Adjust the pre-trained model according to the target diagnosis requirements, the model feature information, the training set, and the validation set to obtain a target model;

[0011] Construct a system according to the target model to obtain a target diagnosis system;

[0012] Diagnose the target vehicle according to the target diagnosis system to obtain a target diagnosis result.

[0013] Second aspect, an embodiment of the present application provides a vehicle diagnostic device with an intelligent system architecture. The device includes a first acquisition module, a labeling module, a partitioning module, a second acquisition module, an adjustment module, a construction module, and a diagnostic module, where:

[0014] The first acquisition module is configured to acquire the target diagnostic requirements and historical diagnostic data of the target vehicle;

[0015] The labeling module is configured to label the historical diagnostic data according to a preset dictionary format to obtain reference data;

[0016] The partitioning module is configured to partition the reference data into a training set and a validation set according to a preset ratio;

[0017] The second acquisition module is configured to acquire a pre-trained model and the model feature information of the pre-trained model;

[0018] The adjustment module is configured to adjust the pre-trained model according to the target diagnostic requirements, the model feature information, the training set, and the validation set to obtain a target model;

[0019] The construction module is configured to perform system construction according to the target model to obtain a target diagnostic system;

[0020] The diagnostic module is configured to diagnose the target vehicle according to the target diagnostic system to obtain a target diagnostic result.

[0021] Third aspect, an embodiment of the present application provides an electronic device, including a processor, a memory, a communication interface, and one or more programs. Among them, the above one or more programs are stored in the above memory and are configured to be executed by the above processor. The above programs include instructions for executing the steps in any method of the first aspect of the embodiments of the present application.

[0022] Fourth aspect, an embodiment of the present application provides a computer-readable storage medium. The computer-readable storage medium stores a computer program for electronic data exchange. The computer program enables a computer to execute some or all of the steps described in any method of the first aspect of the embodiments of the present application.

[0023] Fifth aspect, an embodiment of the present application provides a computer program product. The computer program product includes a non-transitory computer-readable storage medium storing a computer program. The computer program is operable to cause a computer to execute some or all of the steps described in any method of the first aspect of the embodiments of the present application. The computer program product can be a software installation package.

[0024] By implementing the embodiments of the present application, the pre-trained model is adjusted specifically, and a target diagnostic system is constructed, which can accurately and efficiently diagnose vehicles, obtain reliable diagnostic results, realize the intelligence and precision of vehicle diagnosis, reduce the diagnostic cost and improve the diagnostic efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] To more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0026] Figure 1 is the system architecture diagram of a target diagnostic system provided by an embodiment of the present application;

[0027] Figure 2 is the structural schematic diagram of an electronic device provided by an embodiment of the present application;

[0028] Figure 3 is the flowchart of a vehicle diagnosis method for an intelligent system architecture provided by an embodiment of the present application;

[0029] Figure 4 is the flowchart of a data annotation provided by an embodiment of the present application;

[0030] Figure 5 is the flowchart of a model adjustment provided by an embodiment of the present application;

[0031] Figure 6 is the construction flowchart of a large model layer provided by an embodiment of the present application;

[0032] Figure 7 is the scene architecture diagram of a vehicle diagnosis provided by an embodiment of the present application;

[0033] Figure 8 is the block diagram of the functional modules of a vehicle diagnosis device for an intelligent system architecture provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0034] In order to enable those skilled in the art to better understand the solutions of the present application, the following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present application.

[0035] In the description and claims of this application and the above-mentioned drawings, terms such as "first" and "second" are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "comprise" and "have" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally further include steps or units not listed, or may optionally further include other steps or units inherent to these processes, methods, products or devices.

[0036] It should be understood that the term "and / or" herein is merely a description of the association relationship between associated objects, indicating that three relationships may exist. For example, A and / or B may represent the following three cases: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " herein indicates that the associated objects before and after are in an "or" relationship. "Multiple" as used in the embodiments of this application means two or more.

[0037] "At least one (item)" or a similar expression in the embodiments of this application refers to any combination of these items, including any combination of single items (items) or plural items (items), and means one or more, and multiple means two or more. For example, at least one (item) of a, b, or c may represent the following seven cases: a, b, c, a and b, a and c, b and c, a, b, and c. Each of a, b, and c may be an element or a set containing one or more elements.

[0038] "Connection" as used in the embodiments of this application refers to various connection methods such as direct connection or indirect connection to achieve communication between devices, and this application does not make any limitations in this regard.

[0039] Referring to "embodiment" herein means that a specific feature, structure or characteristic described in connection with the embodiment may be included in at least one embodiment of this application. The appearance of this phrase at various positions in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art will explicitly and implicitly understand that the embodiments described herein may be combined with other embodiments.

[0040] The following first explains the relevant terms involved in this application, as follows:

[0041] Agent System: It refers to a software system or entity with characteristics such as intelligence, autonomy, and interactivity. It is usually composed of one or more intelligent Agents, which can be software programs, hardware devices, or other entities with certain intelligent behaviors. Among them, an Agent can perceive information in the environment, reason and make decisions based on its own goals and knowledge, and affect the environment by performing corresponding actions to achieve specific tasks or goals.

[0042] With the development of automotive technologies, especially the rise of new energy vehicles and intelligent connected vehicles, the field of vehicle diagnosis faces many challenges. Due to the rapid technological updates, it is necessary to continuously update equipment and diagnostic technologies to fit the characteristics of new automotive technologies. At the same time, the repair data and diagnostic codes of automobile manufacturers are mostly proprietary, and it is difficult for independent repair shops to obtain the latest repair information. Moreover, although vehicle fault phenomena are similar, the causes are diverse, which will greatly affect the diagnostic accuracy and efficiency. Therefore, how to improve the accuracy and efficiency of vehicle diagnosis urgently needs to be solved.

[0043] To solve the above problems, the embodiments of the present application provide a vehicle diagnosis method and related devices with an intelligent system architecture. First, obtain the target diagnosis requirements and historical diagnosis data of the target vehicle; label the historical diagnosis data according to a preset dictionary format to obtain reference data; divide the reference data into a training set and a validation set according to a preset ratio; obtain a pre-trained model and the model feature information of the pre-trained model; adjust the pre-trained model according to the target diagnosis requirements, the model feature information, the training set, and the validation set to obtain a target model; construct a target diagnosis system according to the target model; and diagnose the target vehicle according to the target diagnosis system to obtain a target diagnosis result. By making targeted adjustments to the pre-trained model and constructing a target diagnosis system, it is possible to accurately and efficiently diagnose vehicles, obtain reliable diagnosis results, realize the intelligence and precision of vehicle diagnosis, reduce the diagnostic cost, and improve the diagnostic efficiency.

[0044] Please refer to Figure 1 , Figure 1 which is the system architecture diagram of a target diagnosis system provided by the embodiments of the present application. The target diagnosis system includes a large model layer, a cache layer, a data layer, a startup layer, and a function layer. Among them, the target diagnosis system is an Agent system that can handle various different types of tasks, such as natural language processing, data analysis, code generation, and automation control. When processing decision-making tasks, it can efficiently evaluate various possibilities and select the best path.

[0045] Among them, the large model layer includes a pre-trained model or other complex models, which can be deployed locally or accessed through the official interface of the large model, and specific limitations are not made here. Among them, the rich knowledge contained in the large model can be used to deeply infer and analyze the relevant information of the target vehicle. For example, based on various phenomenon descriptions and sensor data when the vehicle fails, the possible fault causes and locations can be inferred. The large model layer can automatically extract key features from a large amount of vehicle data, identify potential patterns in the data, and mine complex associations between the data, so as to discover some fault signs or hidden dangers that are difficult to directly detect by traditional methods. Based on the learning and reasoning ability of the model, it provides an intelligent decision-making basis for vehicle diagnosis, gives reasonable diagnosis suggestions and solutions for different vehicle conditions, and provides guidance for subsequent maintenance and repair.

[0046] Among them, the cache layer is used to store frequently accessed data or calculation results. For example, this cache layer is used to store the chat cache information of each user and form a session cache file locally, which can be used to quickly load the historical information of each user, effectively improve the loading speed of historical information, so as to quickly link to the previous chat records, making the target diagnosis system more continuous. Among them, this cache layer can also store the recently used vehicle diagnosis data, inference results of the large model, etc. When the same or similar data is needed again, it can be directly obtained from the cache, greatly reducing the time for data reading and processing and improving the system response speed. The cache layer, as a temporary storage area in the data processing process, facilitates intermediate processing and conversion of data, provides convenience for subsequent analysis and calculation, and helps to improve the fluency and efficiency of data processing.

[0047] Among them, the data layer is responsible for storing and managing various data required by the system, including the historical diagnosis data of the target vehicle, the basic information of the vehicle, sensor data, etc. The data layer will integrate and preprocess data from different channels and in different formats, such as data cleaning, format conversion, standardization, etc., so that the data can meet the input requirements of the large model layer and the function layer, and improve the data quality and usability. The data layer is also responsible for the security management of data. Through technologies such as access control and encryption, it ensures the security and confidentiality of data, and at the same time performs data backup to prevent data loss, ensuring that data can be quickly restored in case of system failure or data corruption.

[0048] It should be noted that the content of the data layer is mainly stored in the Mysql database. The content stored in the tables includes, but is not limited to, the user basic information table, the chat record table, the knowledge base file conversation record table, and the session record table, which are not specifically defined here. Among them, the user basic information table includes the registered user name, password, and user ID of the user; the chat record table includes the chat record ID, session ID, chat type, user question, and model answer; the knowledge base file conversation record table includes the knowledge file ID, file name, name of the affiliated knowledge base, name of the document loader, name of the text splitter, file modification time, file type, file size, number of segmented documents, and creation time; the session record table includes the session ID, user ID, dialog box name, chat type, and creation time, which are not specifically defined here.

[0049] Among them, the startup layer is responsible for the initialization and startup of the system, including operations such as loading necessary configuration files, initializing the large model, and establishing connections with other layers. The startup layer ensures that each component of the system can be correctly loaded and run during startup, preparing for the normal operation of the system. The startup layer can also coordinate and allocate the system's hardware resources, such as CPU, memory, storage, etc., reasonably schedule computing resources, ensure that the large model layer, function layer, etc. can operate efficiently with sufficient resources, and avoid resource conflicts and waste. Among them, the startup layer can start various services required by the system, such as data services, model services, etc., and monitor the startup status of these services in real time, promptly discover and handle problems that occur during the startup process, and ensure the stable startup and operation of the system.

[0050] It should be noted that the startup layer can be started through the FastAPI framework, which is not specifically defined here. This framework can quickly and conveniently provide external interfaces, facilitating other systems or clients to call, promoting the integration and interaction between systems, and enhancing the openness and scalability of the system; it can asynchronously process the operation logic of the project, thus avoiding thread blocking, making full use of system resources, improving the overall performance and response speed of the system, and enabling the system to maintain stable operation under high concurrency; it has the function of managing different processes and can effectively organize and coordinate each component and task within the system; it can start multiple interfaces to access externally at the same time, meeting the call requirements of different clients for different functions of the system. The client includes, but is not limited to, mobile clients, network clients, in-vehicle clients, which are not specifically defined here.

[0051] Among them, the functional layer is used to implement various specific functions of the system, such as user management, session management, knowledge base management, database management, and tool block management. User management is used to manage and store the registration information, session information, etc. of users into the database, and save the relevant information as various data tables in the data layer for storage; Session management is used to manage user sessions, including session creation, maintenance, termination, etc., can track the user's operation process and interaction history, save session status information and manage and store it in the database; Knowledge base management is used to maintain and update the knowledge base of the system, including operations such as knowledge entry, review, classification, retrieval, etc., and import it into the vector database; Database management is used to manage the vector database and the Mysql database, including operations such as database storage, backup, recovery, optimization, etc. When the user selects a different vector database, record this process and store the record in the Mysql database; Tool block management is used to add the tool classes used by the management Agent. Among them, the tool classes include, but are not limited to, Optical Character Recognition (OCR) functions, text processing functions, and search engine modules, which can automatically select the tools suitable for the user's questions according to the user's intention, and record the process of each question and answer in the Mysql database.

[0052] It can be seen that by constructing the target diagnostic system, the ability of automated processing and natural language interaction is provided for users, which can efficiently and accurately troubleshoot faults with the system, improve work efficiency and user experience, realize the intelligence and precision of vehicle diagnosis, reduce diagnostic costs and improve diagnostic efficiency.

[0053] The following Figure 2 is used to describe the electronic device in the embodiments of the present application. Figure 2 is a schematic structural diagram of an electronic device provided by an embodiment of the present application. As Figure 2 shown, the electronic device includes one or more processors, a memory, a communication interface, and one or more programs. The processor is communicatively connected to the memory and the communication interface through an internal communication bus.

[0054] Among them, the processor is mainly used for:

[0055] Obtain the target diagnostic requirements and historical diagnostic data of the target vehicle;

[0056] Label the historical diagnostic data according to a preset dictionary format to obtain reference data;

[0057] Divide the reference data into a training set and a validation set according to a preset ratio;

[0058] Obtain a pre-trained model and the model feature information of the pre-trained model;

[0059] Adjust the pre-trained model according to the target diagnosis requirements, the model feature information, the training set, and the validation set to obtain a target model;

[0060] Construct a system according to the target model to obtain a target diagnosis system;

[0061] Diagnose the target vehicle according to the target diagnosis system to obtain a target diagnosis result.

[0062] Wherein, the one or more programs are stored in the above-mentioned memory and are configured to be executed by the above-mentioned processor. The one or more programs include instructions for executing any step in the above-mentioned method embodiments.

[0063] Wherein, the processor can be, for example, a Central Processing Unit (CPU), a general-purpose processor, a Digital Signal Processor (DSP), an Application-Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute various exemplary logic blocks, units, and circuits described in connection with the disclosure of the present application. The processor can also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, and so on. The communication unit can be a communication interface, a transceiver, a transceiver circuit, etc., and the storage unit can be a memory.

[0064] The memory can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory can be a random access memory (RAM), which is used as an external cache. By way of example but not limitation, many forms of random access memory (RAM) are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchlink dynamic random access memory (SLDRAM), and direct rambus random access memory (DR RAM).

[0065] It can be understood that the electronic device may include more or fewer structural elements than those in the above structural block diagram. For example, it includes a power module, physical buttons, a Wi-Fi module, a speaker, a Bluetooth module, sensors, a display module, etc., which are not limited herein. It can be understood that the electronic device can carry the Figure 1 system architecture as described above.

[0066] After understanding the software and hardware architecture of the present application, the following will describe a vehicle diagnosis method for an intelligent system architecture in an embodiment of the present application in combination with Figure 3 FIG. Figure 3 is a schematic flowchart of a vehicle diagnosis method for an intelligent system architecture provided by an embodiment of the present application, which specifically includes the following steps:

[0067] Step S301, obtain the target diagnosis requirement and historical diagnosis data of the target vehicle.

[0068] Specifically, the target diagnostic requirements of the user for the target vehicle can be obtained through the client, such as the fault phenomena occurring in the target vehicle, the components or systems that need to be detected in the target vehicle, etc.; or the default target diagnostic requirements can be automatically generated according to information such as the type, model, and service life of the target vehicle. For example, for a vehicle with a long service life, the system may automatically take the detection of engine aging, component wear, etc. as the target diagnostic requirements; it can also be based on the real-time operation data, historical fault records of the target vehicle, and the intelligent analysis of the large model layer, and the system automatically identifies possible problems and diagnostic items that need to be carried out. For example, when the fuel consumption of the target vehicle suddenly increases, the system will list the fuel system, engine combustion conditions, etc. as the target diagnostic requirements. In addition, the historical diagnostic data such as the historical diagnostic reports, fault records, and detection data of the target vehicle can be retrieved from the database, or the recent historical diagnostic data of the target vehicle can be quickly loaded through the stored historical cache files, which are not specifically limited here.

[0069] It can be seen that by clarifying the target diagnostic requirements and historical diagnostic data, unnecessary or repetitive labor can be avoided, thereby saving the diagnostic time of the vehicle and improving the diagnostic efficiency.

[0070] Step S302: Label the historical diagnostic data according to a preset dictionary format to obtain reference data.

[0071] For easy understanding, please refer to Figure 4 , Figure 4 which is a schematic flow diagram of data annotation provided by an embodiment of the present application. The specific steps include:

[0072] A1. Determine the first data and the second data in the historical diagnostic data; the data type of the first data is structured data; the data type of the second data is unstructured data;

[0073] A2. Label the first data according to the dictionary format to obtain the first reference data;

[0074] A3. Extract the text from the second data to obtain the text content;

[0075] A4. Integrate the text content to obtain the third data; the data type of the third data is the structured data;

[0076] A5. Label the third data according to the dictionary format to obtain the second reference data;

[0077] A6. Determine the reference data according to the first reference data and the second reference data.

[0078] In a specific embodiment, first, historical diagnostic data is divided into first data and second data. The data type of the first data is structured data, for example, vehicle fault codes, sensor values, etc. stored in tabular form. The data type of the second data is unstructured data, for example, text descriptions of repair records, text information feedback by users, relevant images of vehicle diagnosis, etc. Then, the first data is labeled according to a preset dictionary format to obtain first reference data.

[0079] Next, text extraction is performed on the second data to obtain text content. Among them, the text content corresponding to the second data can be extracted through OCR technology, and specific limitations are not provided here. Then, the text content is integrated into structured third data. Among them, key entity information can be extracted from the text content using natural language processing technology, and then, according to a predefined category or template, the extracted key entity information is classified. Then, the information under each classification is labeled to clarify its specific fields. Finally, the classified and labeled information is organized into structured third data.

[0080] Then, the third data is labeled according to the dictionary format to obtain second reference data. Finally, the reference data is determined based on the first reference data and the second reference data. It should be noted that the dictionary format refers to a structure or rule for standardizing and normalizing data annotation. Data can be stored in the form of key-value pairs, where each key corresponds to a specific data item and the value is the specific content of that data item; it can also be stored in tabular form, including multiple columns and rows, where each column represents a specific data attribute and each row corresponds to a specific data instance; it can also be stored in a tree-like or hierarchical structure. For example, in vehicle diagnosis, each system of the target vehicle can be used as the top-level node, and each system contains multiple subsystems and components below, and each component corresponds to relevant diagnostic data, and specific limitations are not provided here.

[0081] It can be seen that through classification processing and annotation, different types of data are converted into a unified and standardized format, which not only improves the quality and usability of the data but also enhances the retrieval speed and efficiency of the data.

[0082] Step S303, divide the reference data into a training set and a validation set according to a preset ratio.

[0083] Specifically, the preset ratio can be 7:3 or 8:2, and no specific limitation is made here. First, the corresponding preset ratio can be selected according to the size of the data volume and the complexity of the problem. If the data volume is small, the ratio of the training set can be appropriately increased to ensure that the model can learn sufficient features; if the data volume is large, the training effect of the model and the effectiveness of verification can still be ensured after appropriately reducing the ratio of the training set. Then, the reference data is randomly divided into a training set and a verification set according to the preset ratio. For example, 70% of the reference data is the training set, and 30% of the reference data is the verification set. It should be noted that to ensure the model performance of the model, the reference data can also be divided into a training set, a verification set and a test set according to a specific ratio, and after the model training and verification are completed, the model performance of the model is finally evaluated through the test set.

[0084] It can be seen that by reasonably dividing the reference data into a training set and a verification set, a reliable data basis can be provided for the training and evaluation of the model, helping to select a suitable model and adjust model parameters, and improving the performance and generalization ability of the model in practical applications.

[0085] Step S304, obtain a pre-trained model and the model feature information of the pre-trained model.

[0086] Specifically, the corresponding pre-trained model can be obtained through an open-source model library or a model provider. Then, the model feature information can be determined through the relevant documents of the pre-trained model. For example, the architecture of the model, the format requirements of the input and output, the parameter type, etc. are not specifically limited here.

[0087] Step S305, adjust the pre-trained model according to the target diagnosis requirement, the model feature information, the training set and the verification set to obtain a target model.

[0088] For easy understanding, please refer to Figure 5 , Figure 5 which is a schematic flowchart of a model adjustment provided by an embodiment of the present application. The specific steps include:

[0089] B1. Determine initial parameters according to the target diagnosis requirement and the model feature information;

[0090] B2. Configure the pre-trained model according to the initial parameters to obtain a first reference model;

[0091] B3. Fine-tune and train the first reference model according to the training set to obtain a second reference model;

[0092] B4. Evaluate the second reference model according to the verification set to obtain an evaluation result;

[0093] B5. Adjust the second reference model according to the evaluation result to obtain the target model.

[0094] In a specific embodiment, first, analyze the target diagnosis requirement and model feature information to determine the initial parameters of the pre-trained model. Then, configure the pre-trained model according to the initial parameters to obtain the first reference model. Next, fine-tune the first reference model according to the training set to obtain the second reference model. Evaluate the second reference model according to the validation set to obtain the evaluation result. Finally, adjust the second reference model according to the evaluation result to obtain the target model.

[0095] Among them, the specific steps of determining the initial parameters according to the target diagnosis requirement and the model feature information include:

[0096] C1. Determine the diagnosis task type and diagnosis task difficulty corresponding to the target diagnosis requirement;

[0097] C2. Determine the basic structure and initial performance of the pre-trained model according to the model feature information;

[0098] C3. Determine the partial structure corresponding to the diagnosis task type in the basic structure to obtain the reference structure;

[0099] C4. Determine the first parameter corresponding to the reference structure;

[0100] C5. Determine the second parameter of the pre-trained model according to the diagnosis task difficulty and the initial performance;

[0101] C6. Determine the initial parameter according to the first parameter and the second parameter.

[0102] In a specific embodiment, first, determine the diagnosis task type and diagnosis task difficulty corresponding to the target diagnosis requirement. The diagnosis task type includes but is not limited to fault classification, fault prediction, and fault cause analysis, and no specific limitation is made here. The diagnosis task difficulty is comprehensively determined by various factors. For example, the complexity of the data, the abstraction level of the task, the amount of information available, etc. In the case of rare vehicle faults, since there is less relevant data, the diagnosis task difficulty will be greater. Then, determine the basic structure and initial performance of the pre-trained model according to the model feature information. Among them, the model feature information includes the basic structure of the pre-trained model, such as the number of layers included in the neural network model, the number of neurons in each layer, and the connection method between layers. The initial performance can be determined by viewing the evaluation metrics used by the pre-trained model in the pre-training stage, such as accuracy, recall rate, mean square error, etc. For example, if the accuracy of a pre-trained text classification model on its pre-training dataset is 80%, then this result is its initial performance.

[0103] Next, determine the partial structure corresponding to the diagnostic task type in the basic structure to obtain a reference structure. Among them, different diagnostic task types may only be related to a part of the basic structure of the pre-trained model. For example, for the vehicle fault classification task, if the pre-trained model is a general deep learning model, only the output layer and some of the directly connected hidden layers need to be concerned, and this part is mainly responsible for making classification decisions on the input information. Extract the structure corresponding to this part and the diagnostic task type to obtain the reference structure. Then determine the first parameter corresponding to the reference structure. The first parameter can be the connection weights between neurons, biases, etc., which are not specifically limited here.

[0104] Finally, determine the second parameter of the pre-trained model according to the diagnostic task difficulty and the initial performance. Among them, the second parameter includes hyperparameters related to the training process, such as the learning rate, the number of training rounds, the batch size, etc. If the diagnostic task is more difficult and the initial performance of the pre-trained model is average, it may be necessary to appropriately reduce the learning rate and increase the number of training rounds to allow the model to have more time and a more stable learning process to adapt to the task. For example, for a complex vehicle fault prediction task, the learning rate may be set to a smaller value, such as 0.001, and the number of training rounds may be increased to 100 rounds. Then combine the first parameter and the second parameter to obtain the initial parameter.

[0105] It can be seen that by comprehensively considering the target diagnostic requirements and the characteristics of the pre-trained model, the initial parameters corresponding to the automotive diagnostic field are scientifically and reasonably determined, laying a foundation for the subsequent optimization and accurate diagnosis of the model.

[0106] Among them, the training set includes n training subsets, where n is an integer greater than 1. The step of fine-tuning and training the first reference model according to the training set to obtain a second reference model specifically includes:

[0107] D1. Obtain a first training subset; the first training subset is any one of the n training subsets;

[0108] D2. Input the first training subset into the first reference model to obtain a first output result;

[0109] D3. Determine a first loss function according to the diagnostic task type;

[0110] D4. Determine a first true result corresponding to the first output result;

[0111] D5. Calculate the loss value between the first output result and the first true result according to the first loss function to obtain a first loss value;

[0112] D6. Determine a first gradient corresponding to the first loss value;

[0113] D7. Adjust the initial parameters according to the first gradient to obtain first fine-tuning parameters;

[0114] D8. Fine-tune the first reference model according to the first fine-tuning parameters to obtain a first fine-tuned model;

[0115] D9. Determine whether the first fine-tuned model meets a preset condition;

[0116] D10. If the first fine-tuned model meets the preset condition, determine the first fine-tuned model as the second reference model;

[0117] D11. If the first fine-tuned model does not meet the preset condition, continue to fine-tune and train the first fine-tuned model according to the n - 1 training subsets other than the first training subset in the n training subsets to obtain the second reference model.

[0118] In a specific embodiment, first, to improve the training efficiency and effect, the training set is divided into n training subsets, and then one of them is randomly selected as the first training subset, which is convenient for training the first reference model batch by batch according to each training subset, so that the first reference model can gradually learn the features and rules in the data. Then, the data in the first training subset is input into the first reference model, and the first reference model will calculate and process the input data according to its internal parameters and structure, and finally output a prediction result, that is, a first output result. For example, in a vehicle fault classification task, after inputting the fault feature data of the vehicle, the first reference model outputs a prediction result of the fault type.

[0119] Next, determine a first loss function according to the diagnostic task type. Among them, different diagnostic task types need to use different loss functions. For example, for a classification task, such as judging the vehicle fault type, the corresponding first loss function can be a cross-entropy loss function; for a regression task, such as predicting the time when a vehicle fault occurs, the corresponding first loss function can be a mean squared error loss function. Then, determine a first true result corresponding to the first output result. Then, substitute the first output result and the first true result into the first loss function for calculation to obtain a first loss value. The first loss value reflects the prediction error size of the first reference model on the first training subset. The smaller the first loss value, the closer the prediction result of the first reference model is to the true result, and the better the performance of the first reference model.

[0120] Then, calculate the first gradient corresponding to the first loss value. The first gradient represents the rate of change and the direction of change of the first loss function under the initial parameters. Then, according to the calculated first gradient, use a preset optimization algorithm to adjust the initial parameters to obtain the first fine-tuned parameters. Among them, the optimization algorithm will update the initial parameters of the first reference model with a specific step size according to the direction and magnitude of the first gradient, so as to obtain the first fine-tuned parameters. For example, when the optimization algorithm is the stochastic gradient descent algorithm, the initial parameters can be subtracted by the product of the first gradient and the learning rate, so that the first loss value gradually decreases. Apply the first fine-tuned parameters to the first reference model to update the parameters of the first reference model, so as to obtain the first fine-tuned model.

[0121] Finally, determine whether the first fine-tuned model meets the preset conditions. The preset conditions can be set according to actual needs. For example, the loss value is less than a specific threshold, the accuracy rate reaches a specific level, etc., which are not specifically limited here. If the first fine-tuned model meets the preset conditions, it means that the first fine-tuned model has achieved the expected performance under the current training, and it can be determined as the second reference model. Subsequently, operations such as continuing to evaluate it using the validation set can be performed. If the first fine-tuned model does not meet the preset conditions, it means that the first fine-tuned model still needs further training. At this time, use the n - 1 training subsets other than the first training subset to continue the fine-tuning training of the first fine-tuned model, and repeat the above steps until the model meets the preset conditions, and finally obtain the second reference model.

[0122] It can be seen that by continuously adjusting the model parameters to make the model gradually adapt to the training data related to vehicle diagnosis, the inference speed and execution efficiency of the model can be improved.

[0123] Among them, the specific steps of adjusting the second reference model according to the evaluation result to obtain the target model include:

[0124] E1. Determine the evaluation index and evaluation value corresponding to the evaluation result;

[0125] E2. Determine the reference value range and the third parameter corresponding to the evaluation index;

[0126] E3. If the evaluation value is within the reference value range, determine the second reference model as the target model;

[0127] E4. If the evaluation value is not within the reference value range, determine the maximum value and the minimum value of the reference value range;

[0128] E5. Calculate the difference between the evaluation value and the maximum value or the minimum value to obtain the first difference;

[0129] E6. Determine the adjustment factor corresponding to the first difference;

[0130] E7. Adjust the third parameter according to the adjustment factor to obtain a fourth parameter;

[0131] E8. Adjust the second reference model according to the fourth parameter to obtain the target model.

[0132] In a specific embodiment, first, after evaluating the second reference model using a validation set, the corresponding evaluation results are obtained, and then the evaluation metrics and evaluation values corresponding to the evaluation results are determined. Among them, the evaluation metric is a specific criterion for measuring the performance of the model. For example, in the classification task of vehicle fault diagnosis, the evaluation metric can be accuracy, which is the ratio of the number of correctly predicted samples to the total number of samples, or recall rate, which represents the proportion of samples that are actually positive and are correctly predicted as positive. No specific limitation is made here. The evaluation value is the specific value calculated for the evaluation metric by the second reference model on the validation set, such as the accuracy being 0.85, etc. Then, the reference value range and the third parameter corresponding to the evaluation metric are determined. Among them, the reference value range can be determined according to actual experience, professional domain knowledge, or previous experimental results. For example, for the accuracy of a vehicle fault diagnosis model, the corresponding reference value range may be set to 0.8 - 0.95. The third parameter is a hyperparameter related to the model, such as the learning rate, regularization parameter, etc. The third parameter will affect the training and performance of the second reference model.

[0133] Next, when the evaluation value is within the reference value range, it indicates that the performance of the second reference model has reached the expected standard and no further adjustment is required, so the second reference model can be determined as the final target model. When the evaluation value is not within the reference value range, it indicates that the performance of the second reference model has not reached the ideal state and the second reference model needs to be adjusted. Then, the maximum and minimum values of the reference value range are determined. Then, calculate the difference between the evaluation value and the maximum or minimum value to obtain the first difference. If the evaluation value is less than the minimum value, calculate the first difference between the evaluation value and the minimum value; if the evaluation value is greater than the maximum value, calculate the first difference between the evaluation value and the maximum value. According to the mapping relationship between the first difference and the adjustment factor, determine the adjustment factor corresponding to the first difference. The larger the first difference, the larger the adjustment factor, to adjust the model parameters more significantly. Then, adjust the third parameter according to the adjustment factor to obtain the fourth parameter. Finally, apply the fourth parameter to the second reference model to make corresponding adjustments to the second reference model, thereby obtaining the target model.

[0134] It can be seen that according to the evaluation results of the model on the validation set, the model is adjusted and optimized targeted, so as to obtain a target model with better performance, and improve the diagnostic accuracy and reliability of the model in practical applications.

[0135] Step S306, construct a system according to the target model to obtain a target diagnostic system.

[0136] Among them, the specific steps of constructing a system according to the target model to obtain a target diagnostic system include:

[0137] F1. Determine the hardware facilities corresponding to the target model;

[0138] F2. Determine the deep learning framework compatible with the hardware facilities;

[0139] F3. Deploy the target model according to the deep learning framework to obtain a large model layer;

[0140] F4. Integrate the reference data to obtain a data layer;

[0141] F5. Obtain a preset cache layer, startup layer and function layer; the cache layer is used to store cache information; the startup layer is used to provide system interfaces; the function layer includes a user management module, a session management module, a knowledge base management module, a database management module and a tool block management module;

[0142] F6. Construct a system according to the large model layer, the data layer, the cache layer, the startup layer and the function layer to obtain the target diagnostic system.

[0143] In a specific embodiment, first, determine the corresponding hardware facilities according to the target model. Among them, different types and scales of target models require different hardware resources, so their corresponding hardware facilities are different. Then, select a deep learning framework compatible with the hardware facilities, and then deploy the target model to the corresponding hardware facilities through the deep learning framework to obtain a large model layer. Among them, the process of model deployment includes loading information such as the parameters and structure of the target model into the hardware facilities, and performing necessary configuration and optimization so that it can efficiently perform inference and calculation in the actual operating environment. Then, integrate the reference data, and reasonably organize and index the reference data to form a data layer, so that the large model layer can quickly and accurately read and use the data.

[0144] Next, obtain the pre-set cache layer, startup layer, and function layer. Among them, the cache layer is used to store cache information, such as the chat cache information of users, the intermediate calculation results of models, etc. The cache layer can improve the response speed of the system, reduce the repeated calculation and reading of data. For example, cache the chat records between users and the system. When the user requests relevant information again, it can be directly obtained from the cache layer without having to query the database again or perform complex calculations. The startup layer is used to provide system interfaces, which can be implemented through tools such as the FastAPI framework. The startup layer is responsible for functions such as system initialization, resource allocation, and scheduling, and can provide various interfaces externally to facilitate other systems or users to interact with the target diagnostic system. For example, through the interfaces provided by the startup layer, users can send vehicle diagnostic requests and obtain diagnostic results, etc. The function layer includes a user management module, a session management module, a knowledge base management module, a database management module, and a tool block management module. The user management module is responsible for user registration, login, permission management, etc.; the session management module manages the sessions between users and the system; the knowledge base management module maintains the knowledge base of the system; the database management module manages the data layer; the tool block management module manages various intelligent tool blocks in the system. Each functional module of the function layer works together to achieve various specific functions of the system.

[0145] Finally, integrate and connect the large model layer, data layer, cache layer, startup layer, and function layer to build a complete target diagnostic system. The layers interact with each other through interfaces and data transmission. The large model layer obtains data from the data layer for reasoning and calculation, and returns the results to the user or other modules; the cache layer provides data caching and acceleration functions for the system; the startup layer is responsible for system startup and interface provision; the function layer implements various specific functions of the system. Through the cooperation of each layer, a target diagnostic system with complete functions and efficient operation is formed to meet the needs of practical applications such as vehicle diagnosis.

[0146] It can be seen that by building a multi-modal target diagnostic system, not only can tasks such as vehicle diagnosis be accurately performed, but it also has good scalability and maintainability, can adapt to changing business requirements and data characteristics, and provides reliable support for practical applications.

[0147] Step S307, diagnose the target vehicle according to the target diagnostic system to obtain a target diagnostic result.

[0148] Specifically, the target diagnostic system first obtains relevant data from the target vehicle. The relevant data includes the operation data transmitted in real time by various sensors of the target vehicle, such as the operation data corresponding to the engine temperature sensor, speed sensor, oil pressure sensor, etc., as well as the vehicle's historical maintenance records, maintenance information, etc. After the relevant data is collected, it undergoes appropriate preprocessing, such as data cleaning, format conversion, etc., and is input into the data layer of the target diagnostic system for storage and management. For example, the sensor data of the target vehicle is sent to the target diagnostic system in real time at a certain frequency, and the target diagnostic system checks the integrity and accuracy of the sensor data and stores it in the database after removing outliers. Then, the relevant data in the data layer is transmitted to the large model layer. The target model in the large model layer is obtained through training, adjustment, and deployment, and this target model has the ability to analyze and diagnose vehicle faults. The target model can perform reasoning calculations based on the input data using the features and patterns it has learned. For example, for the input engine operation data, the target model will compare it with the normal operation mode and fault mode learned during the training process to determine whether there are abnormalities and the possible types of faults.

[0149] For ease of understanding, please refer to Figure 6 , Figure 6 is a flowchart for constructing the large model layer provided in an embodiment of the present application. It can be seen that first, a pre-trained model is selected. This pre-trained model is a model pre-trained on a large-scale dataset, such as models like Baichuan, Qwen, ChatGLM, etc., which are not specifically limited here. Then, a dataset related to a specific task is collected and processed to obtain a new task dataset. Next, appropriate fine-tuning parameters, such as the learning rate, batch size, number of training epochs, etc., are set according to the task characteristics and model features. Then, the pre-trained model is further trained on the new task dataset to optimize the model's performance on the new task by adjusting the model weights and parameters. Next, the fine-tuned model is evaluated using the validation set corresponding to the new task dataset, and the model structure and parameters are adjusted according to the evaluation results until satisfactory performance is achieved. Finally, the fine-tuned model is deployed to the actual application scenario, that is, the large model layer in the target diagnostic system, to realize the practical value of the model. This large model layer can perform knowledge content Q&A in the field of automotive diagnosis, and this large model layer can be privately deployed, which can avoid uploading sensitive data to the cloud, thus ensuring data security and privacy.

[0150] Next, during the inference process in the large model layer, each functional module in the functional layer will provide corresponding support. For example, the user management module can ensure that only authorized users can access the diagnostic results, thus guaranteeing the security and privacy of the data. The session management module can maintain the interaction session between the user and the system, record the user's operation and request history, and facilitate the user to view and track the diagnostic process at any time. The knowledge base management module can provide relevant knowledge and experience support for the large model layer. For example, when the target model encounters uncertain situations during the diagnosis process, it can obtain knowledge about the vehicle model, fault type, etc. from the knowledge base to assist the target model in making more accurate judgments. The database management module is responsible for managing and maintaining the data in the data layer, ensuring the reliable storage and rapid retrieval of data, and providing accurate data support for the large model layer and other modules. The tool block management module can provide various data analysis and processing tools to help the large model layer process data more effectively. For example, using data visualization tools to display vehicle data in the form of charts, which is convenient for users to intuitively understand the operating conditions of the target vehicle. Among them, if similar data or calculation results have been processed in previous diagnoses, the relevant information will be stored in the cache layer. When the same or similar situations are encountered again, the system can directly obtain the relevant information from the cache layer, avoid repeated calculations, and speed up the diagnosis process.

[0151] Finally, after the large model layer performs inference calculations, it will generate the target diagnostic result for the target vehicle. The target diagnostic result may include whether the target vehicle has a fault, the type of the fault, the severity of the fault, the possible causes of the fault, and corresponding repair suggestions, etc., which are not specifically defined here. The target diagnostic result will be fed back to the user through the interface provided by the startup layer, and the user can view the detailed diagnostic report through the client.

[0152] It can be seen that through the collaborative work of each layer of the target diagnostic system, from data acquisition to the output of the final diagnostic result, a comprehensive, accurate, and efficient diagnosis of the target vehicle is achieved, providing strong support for vehicle maintenance and repair.

[0153] For ease of understanding, please refer to Figure 7 , Figure 7It is a scenario architecture diagram of vehicle diagnosis provided by an embodiment of the present application. It can be seen that after the target diagnosis system obtains relevant data from the target vehicle, it processes the relevant data, and deeply analyzes and reasons the processed data to obtain the target diagnosis result, and then feeds back the target diagnosis result to the client. The user can view it through this client to clearly understand the diagnosis status of the target vehicle. Among them, the target diagnosis system can save and share the diagnosis result, which is convenient for communication and discussion with other professionals. It can also view the historical diagnosis records, compare the conditions of the target vehicle at different times, and better understand the health change trend of the target vehicle. The actual situation, such as the effect after maintenance, can be fed back to the target diagnosis system through the client, so that the target diagnosis system can continuously optimize the diagnosis model and service quality.

[0154] The above mainly introduces the solution of the embodiment of the present application from the perspective of the execution process on the method side. It can be understood that in order for the electronic device to implement the above functions, it includes the corresponding hardware structure and / or software module for executing each function. Those skilled in the art should easily realize that, combining the units and algorithm steps of each example described in the embodiments provided in this article, the present application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a certain function is executed in the way of hardware or computer software driving hardware depends on the specific application and design constraint conditions of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present application.

[0155] The embodiment of the present application can divide the functional units of the electronic device according to the above method examples. For example, each functional unit can be divided corresponding to each function, or two or more functions can be integrated into one processing unit. The above integrated unit can be implemented in the form of hardware or in the form of a software functional unit. It should be noted that the division of units in the embodiment of the present application is illustrative, only a logical function division, and there may be other division methods in actual implementation.

[0156] In the case of dividing each functional module corresponding to each function, Figure 8 It is a block diagram of the functional module composition of a vehicle diagnosis device with an intelligent system architecture provided by an embodiment of the present application. The vehicle diagnosis device 800 with the intelligent system architecture includes a first acquisition module 810, a marking module 820, a division module 830, a second acquisition module 840, an adjustment module 850, a construction module 860, and a diagnosis module 870, where:

[0157] The first acquisition module 810 is used to acquire the target diagnosis requirement and historical diagnosis data of the target vehicle;

[0158] The annotation module 820 is configured to annotate the historical diagnostic data according to a preset dictionary format to obtain reference data;

[0159] The partitioning module 830 is configured to partition the reference data into a training set and a validation set according to a preset ratio;

[0160] The second acquisition module 840 is configured to acquire a pre-trained model and model feature information of the pre-trained model;

[0161] The adjustment module 850 is configured to adjust the pre-trained model according to the target diagnostic requirement, the model feature information, the training set, and the validation set to obtain a target model;

[0162] The construction module 860 is configured to construct a system according to the target model to obtain a target diagnostic system;

[0163] The diagnostic module 870 is configured to diagnose the target vehicle according to the target diagnostic system to obtain a target diagnostic result.

[0164] Optionally, in terms of annotating the historical diagnostic data according to a preset dictionary format to obtain reference data, the annotation module 820 is specifically configured to:

[0165] Determine first data and second data in the historical diagnostic data; the data type of the first data is structured data; the data type of the second data is unstructured data;

[0166] Annotate the first data according to the dictionary format to obtain first reference data;

[0167] Extract text from the second data to obtain text content;

[0168] Integrate the text content to obtain third data; the data type of the third data is the structured data;

[0169] Annotate the third data according to the dictionary format to obtain second reference data;

[0170] Determine the reference data according to the first reference data and the second reference data.

[0171] Optionally, in terms of adjusting the pre-trained model according to the target diagnostic requirement, the model feature information, the training set, and the validation set to obtain a target model, the adjustment module 850 is specifically configured to:

[0172] Determine initial parameters according to the target diagnostic requirement and the model feature information;

[0173] Configure the pre-trained model according to the initial parameters to obtain a first reference model;

[0174] Fine-tune the first reference model according to the training set to obtain a second reference model;

[0175] Evaluate the second reference model according to the validation set to obtain an evaluation result;

[0176] Adjust the second reference model according to the evaluation result to obtain the target model.

[0177] Optionally, in determining the initial parameters according to the target diagnosis requirement and the model feature information, the adjustment module 850 is further specifically configured to:

[0178] Determine the diagnosis task type and diagnosis task difficulty corresponding to the target diagnosis requirement;

[0179] Determine the basic structure and initial performance of the pre-trained model according to the model feature information;

[0180] Determine the partial structure corresponding to the diagnosis task type in the basic structure to obtain a reference structure;

[0181] Determine the first parameter corresponding to the reference structure;

[0182] Determine the second parameter of the pre-trained model according to the diagnosis task difficulty and the initial performance;

[0183] Determine the initial parameters according to the first parameter and the second parameter.

[0184] Optionally, the training set includes n training subsets, where n is an integer greater than 1. In fine-tuning the first reference model according to the training set to obtain a second reference model, the adjustment module 850 is further specifically configured to:

[0185] Obtain a first training subset; the first training subset is any one of the n training subsets;

[0186] Input the first training subset into the first reference model to obtain a first output result;

[0187] Determine a first loss function according to the diagnosis task type;

[0188] Determine a first true result corresponding to the first output result;

[0189] Calculate a loss value between the first output result and the first true result according to the first loss function to obtain a first loss value;

[0190] Determine the first gradient corresponding to the first loss value;

[0191] Adjust the initial parameters according to the first gradient to obtain first fine-tuned parameters;

[0192] Fine-tune the first reference model according to the first fine-tuned parameters to obtain a first fine-tuned model;

[0193] Determine whether the first fine-tuned model meets a preset condition;

[0194] If the first fine-tuned model meets the preset condition, determine the first fine-tuned model as the second reference model;

[0195] If the first fine-tuned model does not meet the preset condition, continue to fine-tune and train the first fine-tuned model according to the n - 1 training subsets other than the first training subset in the n training subsets to obtain the second reference model.

[0196] Optionally, in terms of adjusting the second reference model according to the evaluation result to obtain the target model, the adjustment module 850 is further specifically configured to:

[0197] Determine the evaluation index and evaluation value corresponding to the evaluation result;

[0198] Determine the reference numerical range and the third parameter corresponding to the evaluation index;

[0199] If the evaluation value is within the reference numerical range, determine the second reference model as the target model;

[0200] If the evaluation value is not within the reference numerical range, determine the maximum value and the minimum value of the reference numerical range;

[0201] Calculate the difference between the evaluation value and the maximum value or the minimum value to obtain a first difference;

[0202] Determine the adjustment factor corresponding to the first difference;

[0203] Adjust the third parameter according to the adjustment factor to obtain a fourth parameter;

[0204] Adjust the second reference model according to the fourth parameter to obtain the target model.

[0205] Optionally, in terms of constructing a target diagnostic system according to the target model, the construction module 860 is specifically configured to:

[0206] Determine the hardware facilities corresponding to the target model;

[0207] Determine a deep learning framework compatible with the hardware facilities;

[0208] Deploy the target model according to the deep learning framework to obtain a large model layer;

[0209] Integrate the reference data to obtain a data layer;

[0210] Obtain a preset cache layer, startup layer, and function layer; the cache layer is used to store cache information; the startup layer is used to provide a system interface; the function layer includes a user management module, a session management module, a knowledge base management module, a database management module, and a tool block management module;

[0211] Construct a system according to the large model layer, the data layer, the cache layer, the startup layer, and the function layer to obtain the target diagnostic system.

[0212] It can be seen that by making targeted adjustments to the pre-trained model and constructing the target diagnostic system, the vehicle can be accurately and efficiently diagnosed, reliable diagnostic results can be obtained, the intelligence and accuracy of vehicle diagnosis can be realized, the diagnostic cost can be reduced, and the diagnostic efficiency can be improved.

[0213] It should be noted that the specific implementation of each operation can adopt the corresponding description of the method embodiment shown above. The vehicle diagnostic device 800 with an intelligent system architecture can be used to execute the method embodiments of the present application above, which will not be elaborated here.

[0214] The embodiment of the present application also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program for electronic data exchange, and the computer program enables a computer to execute some or all of the steps of any method recorded in the method embodiment above. The above computer includes an electronic device.

[0215] The embodiment of the present application also provides a computer program product. The above computer program product includes a non-transitory computer-readable storage medium storing a computer program. The above computer program can operate to enable a computer to execute some or all of the steps of any method recorded in the method embodiment above. The computer program product can be a software installation package, and the above computer includes an electronic device.

[0216] It should be noted that, for each of the above embodiments, for the sake of simple description, they are all expressed as a series of action combinations. Those skilled in the art should be aware that the present application is not limited by the described action sequence, because some steps in the embodiments of the present application can be performed in other sequences or simultaneously. Additionally, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions, steps, modules, or units involved are not necessarily essential to the embodiments of the present application.

[0217] In the above embodiments, the descriptions of the embodiments of the present application each have their own focuses. For parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0218] Those of ordinary skill in the art can understand all or part of the processes of implementing the methods in the above embodiments. These processes can be completed by hardware instructed by a computer program. The program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the above method embodiments. The aforementioned storage medium includes: ROM or random access memory RAM, magnetic disks, or optical discs, etc., which are various media that can store program codes.

[0219] The steps of the methods or algorithms described in the embodiments of the present application can be implemented in a hardware manner or by a processor executing software instructions. The software instructions can be composed of corresponding software modules. The software modules can be stored in RAM, flash memory, ROM, EPROM, electrically erasable programmable read-only memory (EEPROM), registers, hard disks, removable hard disks, compact disc read-only memory (CD-ROM), or any other form of storage medium well-known in the art. An exemplary storage medium is coupled to the processor, enabling the processor to read information from the storage medium and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and the storage medium can be located in an ASIC. Additionally, the ASIC can be located in a terminal device or a management device. Of course, the processor and the storage medium can also exist as discrete components in the terminal device or the management device.

[0220] Those skilled in the art should be able to realize that in one or more of the above examples, the functions described in the embodiments of the present application can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wire (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wirelessly (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that the computer can access or a data storage device such as a server or data center that includes one or more available media integrated. The available medium can be a magnetic medium (such as a floppy disk, hard disk, magnetic tape), an optical medium (such as a digital video disc (DVD)), or a semiconductor medium (such as a solid state disk (SSD)), etc.

[0221] Each device and product described in the above embodiments includes various modules / units, which can be software modules / units, hardware modules / units, or can be partly software modules / units and partly hardware modules / units. For example, for each device and product applied to or integrated into a chip, each of the modules / units it includes can be implemented in the form of hardware such as circuits. Or, at least some of the modules / units can be implemented in the form of software programs that run on the processor integrated inside the chip, and the remaining (if any) part of the modules / units can be implemented in the form of hardware such as circuits. For each device and product applied to or integrated into a chip module, each of the modules / units it includes can be implemented in the form of hardware such as circuits. Different modules / units can be located in the same component (such as a chip, a circuit module, etc.) or different components of the chip module. Or, at least some of the modules / units can be implemented in the form of software programs that run on the processor integrated inside the chip module, and the remaining (if any) part of the modules / units can be implemented in the form of hardware such as circuits. For each device and product applied to or integrated into a terminal device, each of the modules / units it includes can be implemented in the form of hardware such as circuits. Different modules / units can be located in the same component (such as a chip, a circuit module, etc.) or different components inside the terminal device. Or, at least some of the modules / units can be implemented in the form of software programs that run on the processor integrated inside the terminal device, and the remaining (if any) part of the modules / units can be implemented in the form of hardware such as circuits.

[0222] The specific embodiments described above further elaborate on the objectives, technical solutions, and beneficial effects of the embodiments of the present application. It should be understood that the above description is only the specific embodiments of the embodiments of the present application and is not used to limit the protection scope of the embodiments of the present application. Any modifications, equivalent replacements, improvements, etc. made on the basis of the technical solutions of the embodiments of the present application shall be included in the protection scope of the embodiments of the present application.

Claims

1. A vehicle diagnostic method with an intelligent system architecture, characterized in that: The method comprises: Obtain target diagnostic requirements and historical diagnostic data of the target vehicle; Annotating the historical diagnostic data according to a preset dictionary format to obtain reference data; Dividing the reference data into a training set and a validation set according to a preset ratio; Acquire a pre-trained model and model feature information of the pre-trained model; Adjusting the pre-trained model according to the target diagnosis requirement, the model feature information, the training set and the validation set to obtain a target model; Constructing a system according to the target model to obtain a target diagnosis system; The target vehicle is diagnosed according to the target diagnosis system to obtain a target diagnosis result.

2. The method according to claim 1, characterized in that The step of labeling the historical diagnostic data according to a preset dictionary format to obtain reference data includes: Determine first data and second data in the historical diagnostic data; the data type of the first data is structured data; the data type of the second data is unstructured data; Annotating the first data according to the dictionary format to obtain first reference data; Performing text extraction on the second data to obtain text content; Integrate the text content to obtain third data; the data type of the third data is the structured data; Annotating the third data according to the dictionary format to obtain second reference data; The reference data is determined according to the first reference data and the second reference data.

3. The method according to claim 1, characterized in that The step of adjusting the pre-trained model according to the target diagnosis requirement, the model feature information, the training set, and the validation set to obtain a target model includes: Determining initial parameters according to the target diagnosis requirements and the model feature information; Configuring the pre-trained model according to the initial parameters to obtain a first reference model; Fine-tune the first reference model according to the training set to obtain a second reference model; Evaluate the second reference model according to the verification set to obtain an evaluation result; The second reference model is adjusted according to the evaluation result to obtain the target model.

4. The method according to claim 3, characterized in that The determining of initial parameters according to the target diagnosis requirement and the model feature information includes: Determine the diagnostic task type and diagnostic task difficulty corresponding to the target diagnostic requirement; Determining the basic structure and initial performance of the pre-trained model according to the model feature information; Determine a partial structure corresponding to the diagnostic task type in the basic structure to obtain a reference structure; determining a first parameter corresponding to the reference structure; Determining a second parameter of the pre-trained model according to the diagnostic task difficulty and the initial performance; The initial parameter is determined according to the first parameter and the second parameter.

5. The method according to claim 4, characterized in that The training set includes n training subsets, where n is an integer greater than 1, and fine-tuning the first reference model according to the training set to obtain a second reference model includes: Acquire a first training subset; the first training subset is any one of the n training subsets; Inputting the first training subset into the first reference model to obtain a first output result; Determining a first loss function according to the diagnostic task type; Determine a first true result corresponding to the first output result; Calculate the loss value between the first output result and the first true result according to the first loss function to obtain a first loss value; Determine a first gradient corresponding to the first loss value; Adjust the initial parameter according to the first gradient to obtain a first fine-tuning parameter; Fine-tune the first reference model according to the first fine-tuning parameter to obtain a first fine-tuning model; Determining whether the first fine-tuning model meets a preset condition; If the first fine-tuning model meets the preset condition, determining the first fine-tuning model as the second reference model; If the first fine-tuning model does not meet the preset condition, the first fine-tuning model is further fine-tuned according to n-1 training subsets among the n training subsets except the first training subset to obtain the second reference model.

6. The method according to claim 3, characterized in that The step of adjusting the second reference model according to the evaluation result to obtain the target model includes: Determine the evaluation index and evaluation value corresponding to the evaluation result; Determine a reference value range and a third parameter corresponding to the evaluation index; If the evaluation value is within the reference value range, determining the second reference model as the target model; If the evaluation value is not within the reference value range, determining the maximum value and the minimum value of the reference value range; Calculate the difference between the evaluation value and the maximum value or the minimum value to obtain a first difference; determining an adjustment factor corresponding to the first difference; Adjust the third parameter according to the adjustment factor to obtain a fourth parameter; The second reference model is adjusted according to the fourth parameter to obtain the target model.

7. The method according to any one of claims 1 to 6, characterized in that: The system is constructed according to the target model to obtain a target diagnosis system, including: Determining the hardware facilities corresponding to the target model; Determining a deep learning framework that is compatible with the hardware facility; Deploy the target model according to the deep learning framework to obtain a large model layer; Integrating the reference data to obtain a data layer; Obtaining a preset cache layer, a startup layer and a function layer; the cache layer is used to store cache information; the startup layer is used to provide a system interface; the function layer includes a user management module, a session management module, a knowledge base management module, a database management module and a tool block management module; The target diagnosis system is obtained by constructing a system based on the large model layer, the data layer, the cache layer, the startup layer and the functional layer.

8. A vehicle diagnostic device with an intelligent system architecture, characterized in that: The device includes a first acquisition module, a labeling module, a division module, a second acquisition module, an adjustment module, a construction module and a diagnosis module, wherein: The first acquisition module is used to acquire target diagnostic requirements and historical diagnostic data of the target vehicle; The labeling module is used to label the historical diagnostic data according to a preset dictionary format to obtain reference data; The division module is used to divide the reference data into a training set and a validation set according to a preset ratio; The second acquisition module is used to acquire a pre-trained model and model feature information of the pre-trained model; The adjustment module is used to adjust the pre-trained model according to the target diagnosis requirements, the model feature information, the training set and the verification set to obtain a target model; The construction module is used to construct a system according to the target model to obtain a target diagnosis system; The diagnostic module is used to diagnose the target vehicle according to the target diagnostic system to obtain a target diagnostic result.

9. An electronic device, characterized in that: include: A processor, a memory, a communication interface, and one or more programs; The one or more programs are stored in the memory and configured to be executed by the processor, the programs comprising instructions for executing the steps in the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, wherein the computer program includes program instructions, and when the program instructions are executed by a processor, the processor is caused to perform the method according to any one of claims 1 to 7.

Images