Large model upgrading method and device applied to vehicle machine, electronic equipment and readable storage medium

Abstract

Embodiments of the present application provide a large model upgrading method and device applied to a car machine, electronic equipment and a readable storage medium. In some embodiments, the large model upgrading method applied to the car machine comprises the following steps: in the process of loading an original large model on the car machine to perform reasoning based on a prompt word, in response to obtaining a first instruction for confirming to upgrade the original large model, a target large model is acquired and run; in response to determining that the target large model has been successfully run and an intermediate result has been obtained by loading the original large model to perform reasoning, the target large model is loaded to continue to perform reasoning based on the prompt word and the intermediate result, and a first result is obtained.

Description

Technical Field

[0001] The embodiments of this application relate to the field of computer technology, and more specifically, to a method, apparatus, electronic device, and readable storage medium for large-scale model upgrades of vehicle systems. Background Technology

[0002] With the increasing prevalence of large-scale model applications, the installation and running of applications based on large language models on in-vehicle infotainment systems is also becoming more common. Currently, updating large-scale model applications on in-vehicle infotainment systems typically requires a complete system upgrade, with the newly developed large-scale model application packaged into the system's upgrade package for simultaneous upgrade. However, this reduces the efficiency of large-scale model application upgrades and iterations. Summary of the Invention

[0003] The embodiments of this application provide a method, apparatus, electronic device, and readable storage medium for large-scale vehicle system upgrades that can at least partially solve the above-mentioned or other problems existing in the prior art.

[0004] One implementation of this application provides a method for upgrading a large model applied to a vehicle infotainment system, comprising: during the process of loading the original large model on the vehicle infotainment system for reasoning based on prompt words, in response to receiving a first instruction to confirm the upgrade of the original large model, acquiring and running a target large model; in response to determining that the target large model has been successfully run and that intermediate results have been obtained through reasoning by loading the original large model, loading the target large model for further reasoning based on prompt words and intermediate results to obtain a first result.

[0005] Another embodiment of this application provides a large model upgrade device applied to an in-vehicle infotainment system, comprising: an acquisition and operation module configured to, in response to receiving a first instruction confirming the upgrade of the original large model during the process of reasoning based on prompt words using the original large model loaded on the in-vehicle infotainment system, acquire and run a target large model; and a loading and acquisition module configured to, in response to determining that the target large model has been successfully run and that intermediate results have been obtained through reasoning using the original large model, load the target large model to continue reasoning based on prompt words and intermediate results to obtain a first result.

[0006] Another embodiment of this application provides an electronic device, including at least one processor and a memory. The memory is communicatively connected to the at least one processor and stores instructions executable by the at least one processor. These instructions are executed by the at least one processor to enable the at least one processor to perform the large-scale model upgrade method for in-vehicle systems mentioned in the above embodiments.

[0007] Another embodiment of this application provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the large-scale model upgrade method for in-vehicle systems mentioned in the above embodiments.

[0008] According to some embodiments of this application, a method, apparatus, electronic device, and readable storage medium for upgrading large models applied to in-vehicle systems are provided. During the process of loading the original large model on the in-vehicle system and performing inference based on prompt words, in response to receiving a first instruction confirming the upgrade of the original large model, the target large model is acquired and run. Then, in response to determining that the target large model has been successfully run and that intermediate results have been obtained through inference by loading the original large model, the target large model is loaded again to continue inference based on the prompt words and intermediate results to obtain a first result. This achieves hot updating of the large model on the in-vehicle system, meaning that the large model is upgraded while the user is using it. The user is unaware of the large model update during use, thus improving the user experience. Furthermore, timely hot updating of the large model improves the efficiency of upgrading large model applications on the in-vehicle system.

[0009] In some implementations, during the process of loading the original large model on the vehicle's infotainment system and performing inference based on prompt words, in response to receiving a first instruction confirming the upgrade of the original large model, the target large model is acquired and run. This includes: loading the original large model and performing inference based on prompt words through the vehicle's large model inference engine; and, in response to receiving the first instruction, simultaneously loading the original large model and performing inference based on prompt words through the large model inference engine while running the target large model through the large model inference engine. Thus, the large model inference engine enables the upgrade of the large model without affecting the user's use of the large model; that is, it enables hot updates of the large model, thereby improving the user experience.

[0010] In some implementations, the target large model is loaded and reasoned further based on prompts and intermediate results to obtain a first result. This includes loading the target large model through the vehicle's large model inference engine, and then reasoning further based on prompts and intermediate results to obtain a first result. This allows for the upgrading of the large model without affecting user access, improving the efficiency of large model upgrades and enhancing the user experience.

[0011] In some implementations, the large model upgrade method applied to in-vehicle infotainment systems further includes: in response to obtaining prompt words through a large model application installed in the in-vehicle infotainment system, inputting the prompt words into the large model inference engine of the in-vehicle infotainment system via the application programming interface (API) of the large model application. Since the large model and the large model application are not in a persistent connection, inputting the prompt words obtained by the large model application into the large model inference engine via the large model application's API allows the large model loaded by the large model inference engine to perform inference based on the prompt words. This achieves real-time input of user-inputted prompt words into the large model inference engine through the large model application's API, improving the inference efficiency of the large model.

[0012] In some implementations, the large model upgrade method applied to in-vehicle infotainment systems further includes: in response to receiving a large model upgrade instruction including upgrade instruction information from the in-vehicle infotainment system, downloading the target large model and storing it in a large model library at the moment indicated by the upgrade instruction information; wherein, in response to receiving a first instruction confirming the upgrade of the original large model, acquiring and running the target large model includes: in response to receiving the first instruction, acquiring and running the target large model from the large model library through the large model inference engine of the in-vehicle infotainment system. Thus, by first storing the target large model in the large model library, and then having the large model inference engine acquire and run the target large model from the large model library, the efficiency of acquiring the target large model is improved.

[0013] In some implementations, the large model upgrade method applied to the vehicle infotainment system further includes: displaying information requesting an upgrade of the original large model when the upgrade instruction indicates that the original large model needs to be upgraded; and obtaining a first instruction based on a confirmation instruction for the information requesting an upgrade of the original large model. Thus, the instruction information for upgrading the original large model is generated only after the user confirms the upgrade information, allowing the user to choose whether to upgrade the original large model according to their actual needs, further improving the user experience.

[0014] In some implementations, the large model upgrade method applied to in-vehicle systems further includes: in response to determining that a first result has been successfully obtained, releasing the original large model through the large model inference engine of the in-vehicle system. This releases the resources occupied by the original large model on the in-vehicle system when the target large model can operate normally, saving storage resources on the in-vehicle system. Attached Figure Description

[0015] Other features, objects, and advantages of this application will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings. Wherein:

[0016] Figure 1 This is a flowchart illustrating a large-scale model upgrade method for in-vehicle systems according to the first embodiment of this application;

[0017] Figure 2This is a flowchart illustrating a large-scale model upgrade method for in-vehicle systems according to the second embodiment of this application;

[0018] Figure 3 This is a schematic diagram of the application process of the large-scale model upgrade method for vehicle infotainment systems according to this application;

[0019] Figure 4 This is a schematic block diagram of a large-scale vehicle system upgrade device according to some embodiments of this application;

[0020] Figure 5 These are schematic block diagrams of electronic devices according to some embodiments of this application. Detailed Implementation

[0021] To better understand this application, various aspects of this application will be described in more detail with reference to the accompanying drawings. It should be understood that these detailed descriptions are merely illustrative of exemplary embodiments of this application and are not intended to limit the scope of this application in any way. The expression "and / or" includes any and all combinations of one or more of the associated listed items.

[0022] It should also be understood that expressions such as "comprising," "including," "having," "containing," and / or "comprising" are open-ended rather than closed-ended expressions in this specification, indicating the presence of the stated features, elements, and / or components, but not excluding the presence of one or more other features, elements, components, and / or combinations thereof. Furthermore, when expressions such as "at least one of..." appear after a list of listed features, they modify the entire list of features, not just individual elements in the list. Additionally, when describing embodiments of this application, the word "may" is used to mean "one or more embodiments of this application." And the term "exemplary" is intended to refer to examples or illustrations.

[0023] Unless otherwise specified, all terms used herein (including engineering and technical terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. It should also be understood that, unless expressly stated herein, terms defined in common dictionaries shall be interpreted as having the meaning consistent with their meaning in the context of the relevant art, and not as having an idealized or overly formalized meaning.

[0024] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. Furthermore, unless explicitly limited or contradicted by the context, the specific steps included in the methods described in this application are not limited to the order in which they are described, but can be performed in any order or in parallel. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0025] Figure 1 This is a flowchart illustrating a large-scale model upgrade method for an in-vehicle infotainment system according to a first embodiment of this application. The large-scale model upgrade method 100 for an in-vehicle infotainment system may include the following steps:

[0026] Step 101: During the process of reasoning based on prompt words using the original large model loaded on the vehicle's infotainment system, in response to the first instruction to confirm the upgrade of the original large model, the target large model is acquired and run.

[0027] In this embodiment, if the vehicle-mounted terminal receives a first instruction to confirm the upgrade of the original large model during the process of reasoning based on the prompt words while loading the original large model on the vehicle-mounted terminal, it will acquire and run the target large model.

[0028] In this implementation, the original large model will not cease service during the upgrade process, meaning users can still use the in-vehicle application to call the original large model for inference, generation, and other tasks. It should be noted that the large model, also known as the Large Language Model (LLM), is a deep learning-based generative model that captures linguistic statistical patterns through training on massive amounts of data, possessing natural language understanding and generation capabilities. The prompt words are text instructions input by the user to guide the model in completing tasks; their quality directly affects the accuracy and relevance of the output results.

[0029] If, during the process of a user calling the original large model for inference and generation using the vehicle-mounted application, the vehicle-mounted system receives a first instruction to upgrade the original large model, then it will acquire and run the target large model, which is an upgraded version of the original large model.

[0030] In some embodiments of this application, the training process of the large model is carried out in the cloud. After the cloud completes the training of the large language model, it generates a large language model LLM_X. Due to the limited computing power of the vehicle terminal, the cloud performs quantization processing on the trained large model and generates a quantized large model LLM_X_int4 after the quantization processing is completed.

[0031] For the in-vehicle infotainment chip platform, the large language model generated in the cloud after quantization processing is packaged using the packaging tool provided by the chip platform. The packaging process mainly involves converting the format of the large cloud-generated model LLM_X_int4 so that it can run on the in-vehicle infotainment chip platform, resulting in the packaged large language model LLM_X_int4_qnn.

[0032] Then, the vehicle's infotainment system stores the packaged large language model LLM_X_int4_qn in its local model repository. It's important to note that the vehicle's local model repository monitors the cloud-based updates and upgrade plans for large models. If the repository detects a new upgrade plan for a large model, it retrieves the latest version of the large model (the target model, such as LLM_X_int4_qn) based on the new plan, downloads it, and runs it—essentially installing the target large model.

[0033] The cloud platform first sends the upgrade plan to the vehicle's infotainment system. Upon receiving the plan, the system downloads the packaged large model according to its local policies. If the system doesn't block automatic upgrade package downloads, it will download automatically; otherwise, manual download is required. Furthermore, user participation necessitates confirmation; if users agree to automatic downloads, the system will download the package upon receiving the upgrade plan, and user confirmation is required before the update.

[0034] In some embodiments of this application, the vehicle-mounted system will place the downloaded upgrade package into the local large language model storage file partition, then decompress the upgrade package, delete the original version, and finally replace the original large model (original large model) with the new decompressed version (target large model).

[0035] Step 102: In response to determining that the target large model has been successfully run and that intermediate results have been obtained by loading the original large model, the target large model is loaded to continue reasoning based on the prompt words and intermediate results to obtain the first result.

[0036] In this embodiment, if the vehicle-mounted system determines that the target large model has been successfully run and that intermediate results have been obtained through reasoning from the original large model, then the target large model will be loaded so that it can continue reasoning based on the prompt words and intermediate results to obtain the first result.

[0037] Specifically, users can use the in-vehicle application to invoke the original large model for inference. The in-vehicle system combines the user's input in the application with locally stored prompts, and then inputs the combined content into the original large model to enable inference. During this process, the in-vehicle system receives a first instruction to upgrade the original large model, and according to the first instruction, acquires and runs the target large model. If the target large model has been successfully run (or successfully installed) and the original large model has generated intermediate results through inference, the in-vehicle system will then load the successfully running target large model to continue inference. That is, it loads the target large model so that the target large model can continue inference based on the prompts and intermediate results, thereby obtaining the final inference result, i.e., the first result.

[0038] In some embodiments of this application, the large model can be loaded multiple times during the inference process. That is, while an application calls the large model to generate a service, other applications can call the large model to generate another service, and the two services exist simultaneously.

[0039] In some embodiments of this application, after obtaining the first result using the target large model, the first result is sent to the application that calls the large model. After receiving the first result, the application performs related actions based on the content of the first result, such as calling the interface of other applications, or using TTS (Text To Speech) to convert the text into speech for playback.

[0040] According to some embodiments of this application, during the process of loading the original large model on the vehicle's infotainment system and performing inference based on prompt words, in response to receiving a first instruction confirming the upgrade of the original large model, the target large model is acquired and run; then, in response to determining that the target large model has been successfully run and that intermediate results have been obtained through inference by loading the original large model, the target large model is loaded to continue inference based on prompt words and intermediate results to obtain the first result, thereby realizing hot updates of the large model on the vehicle's infotainment system. The large model is upgraded while the user is using it, and the user is unaware of the large model update during the use of the large model, thereby improving the user experience. Furthermore, by performing hot updates of the large model in a timely manner, the upgrade efficiency of the large model application on the vehicle's infotainment system is improved.

[0041] It should be noted that the acquisition, storage, and application of user personal information involved in the technical solution disclosed herein all comply with relevant laws and regulations and do not violate public order and good morals. It should also be noted that the information in this embodiment was obtained after being authorized by the user (i.e., with the user's consent).

[0042] Figure 2This is a flowchart illustrating a large-scale model upgrade method for in-vehicle systems according to the second embodiment of this application, as shown below. Figure 2 As shown, the large-scale model upgrade method 200 applied to the vehicle system may include the following steps:

[0043] Step 201: Load the original large model through the vehicle's large model inference engine and perform inference based on the prompt words.

[0044] In this embodiment, the vehicle-mounted system loads the original large model through a large model inference engine, enabling the large model to perform inference based on the prompt words.

[0045] It's important to note that the large-scale model inference engine is a computing framework specifically designed for generative AI (Artificial Intelligence) with hundreds of billions of parameters. Its core task is to efficiently deploy trained base models to various hardware environments to achieve low-latency, high-throughput real-time inference services. By optimizing computing resource allocation, compressing model size, and dynamically scheduling requests, it supports large-scale applications in complex scenarios such as natural language processing and multimodal interaction.

[0046] In other words, the vehicle's infotainment system is equipped with a large model inference engine. After the large model application (an application based on the large model) on the vehicle's infotainment system is launched, it will receive the user's input. After the user finishes inputting, the user's input is combined with the locally stored prompt, and then the combined content is input into the large model inference engine so that the large model inference engine loads the original large model and performs inference based on the prompt.

[0047] Step 202: In response to receiving the first instruction, while loading the original large model through the large model inference engine and performing inference based on the prompt words, the target large model is run through the large model inference engine.

[0048] In this embodiment, if the vehicle-mounted system receives a first instruction to upgrade the original large model, it will simultaneously load the original large model and perform inference based on the prompt words through the large model inference engine, and also run the target large model through the large model inference engine. That is, the large model inference engine can run the target large model while loading the original large model for inference. Thus, the large model inference engine enables the upgrading of the large model without affecting the user's use of the large model; in other words, it achieves hot updates of the large model, thereby improving the user experience.

[0049] Step 203: In response to the determination that the target large model has been successfully run and intermediate results have been obtained by loading the original large model, the target large model is loaded through the vehicle's large model inference engine, and inference is continued based on the prompt words and intermediate results to obtain the first result.

[0050] In this embodiment, if the vehicle-mounted system determines that the target large model has been successfully running, it will load the target large model through the large model inference engine. If intermediate results have been obtained through inference using the original large model at the same time, the target large model loaded through the large model inference engine will continue to infer based on the prompt words and intermediate results to obtain the first result.

[0051] Specifically, users can use the in-vehicle application to invoke the original large model for inference. The in-vehicle system combines the user's input in the application with locally stored prompts, and then inputs the combined content into the large model inference engine. After loading the original large model stored locally on the in-vehicle system, the original large model performs inference based on the combined content. During this process, the in-vehicle system receives a first instruction to upgrade the original large model. While loading the original large model through the large model inference engine and performing inference based on the prompts, it also runs the target large model through the large model inference engine. If the target large model has been successfully run (or successfully installed) and the original large model has generated intermediate results through inference, the large model inference engine will then load the successfully running target large model to continue inference. That is, the target large model is loaded through the large model inference engine so that it can continue inference based on the prompts and intermediate results, thereby obtaining the final inference result, i.e., the first result.

[0052] According to some embodiments of this application, the vehicle-mounted system loads the original large model based on prompt words through the vehicle-mounted system's large model inference engine. Then, in response to receiving a first instruction, while loading the original large model based on prompt words through the large model inference engine, the system also runs the target large model. Finally, in response to determining that the target large model has been successfully run and that intermediate results have been obtained by loading the original large model, the system loads the target large model through the vehicle-mounted system's large model inference engine, continues inference based on prompt words and intermediate results, and obtains the first result. Thus, the large model can be upgraded without affecting the user's use of the large model, that is, the large model can be hot-updated through the large model inference engine, thereby improving the user experience.

[0053] The steps of the various methods described above are only for clarity. In implementation, they can be combined into one step or some steps can be split into multiple steps. As long as they include the same logical relationship, they are all within the protection scope of this disclosure. Adding insignificant modifications or introducing insignificant designs to the algorithm or process, but without changing the core design of the algorithm and process, are also within the protection scope of this disclosure.

[0054] In some embodiments of this application, the above-mentioned large model upgrade method applied to vehicle infotainment systems further includes: in response to obtaining a prompt word through a large model application installed in the vehicle infotainment system, inputting the prompt word into the large model inference engine of the vehicle infotainment system through the API (Application Programming Interface) of the large model application.

[0055] In this implementation, if the vehicle-mounted system obtains a prompt word through the large model application installed on the vehicle-mounted system, it will input the obtained prompt into the large model inference engine on the vehicle-mounted system through the API of the large model application, so that the original large model and / or the target large model loaded by the large model inference engine can perform inference based on the prompt word.

[0056] The prompts here are text instructions input by the user to the model, guiding it to complete tasks. Their quality directly affects the accuracy and relevance of the output. Specifically, the user inputs information into the large model application, which is then combined with locally stored prompts to generate the final prompts. These prompts are then fed into the large model inference engine via the application's API. This allows the original large model, loaded through the inference engine, to perform inference based on the prompts, and / or the target large model to perform inference based on the prompts and intermediate results.

[0057] Since the large model and the large model application in this implementation are not in a long-lived connection relationship, the prompt words obtained by the large model application are input into the large model inference engine through the API interface of the large model application. This allows the large model loaded by the large model inference engine to perform inference based on the prompt words. Thus, the prompt words input by the user are input into the large model inference engine in real time through the API interface of the large model application, which improves the inference efficiency of the large model and thus enhances the user experience.

[0058] In some embodiments of this application, the above-described method for upgrading large models applied to vehicle infotainment systems further includes: in response to receiving a large model upgrade instruction including upgrade instruction information from the vehicle infotainment system, downloading the target large model and storing it in a large model library at the time indicated by the upgrade instruction information to acquire the target large model; and in response to receiving a first instruction confirming the upgrade of the original large model, acquiring and running the target large model, including: in response to receiving the first instruction, acquiring and running the target large model from the large model library through the vehicle infotainment system's large model inference engine.

[0059] In this implementation, the vehicle's infotainment system has a local repository. If the system receives a large model upgrade command including upgrade instructions, it will download and store the target large model in the large model repository (i.e., the local repository) when retrieving the target large model. Upon receiving the first command, the system will retrieve and run the target large model from the large model repository using the large model inference engine. Therefore, by first storing the target large model in the large model repository and then having the large model inference engine retrieve and run it, the efficiency of target large model retrieval is improved.

[0060] In some embodiments of this application, the above-mentioned large model upgrade method applied to vehicle infotainment systems further includes: displaying information requesting the upgrade of the original large model when the upgrade instruction information indicates the time for upgrading the original large model; and obtaining a first instruction based on the confirmation instruction for the information requesting the upgrade of the original large model.

[0061] In this implementation, the cloud creates an upgrade plan for the vehicle-mounted large language model in the large language model update management service and formulates an upgrade strategy. The upgrade strategy is usually divided into two parts: small-batch upgrade verification and large-batch upgrade push. The small-batch upgrade verification mainly selects a portion of test vehicles from all vehicles for upgrade verification, mainly to verify whether the new large model can run normally on the vehicle-mounted system. After the small-batch upgrade verification is completed, the large-batch upgrade push begins.

[0062] After receiving the upgrade task pushed by the cloud-based large model upgrade management service, the vehicle's infotainment system will display the request to upgrade the original large model to the user, for example, through the in-vehicle screen. Once the user confirms acceptance of the upgrade, a confirmation command is generated to confirm the request to upgrade the original large model, and the first instruction is derived based on this confirmation command. Therefore, the instruction to upgrade the original large model is only generated after the user confirms the upgrade information. This allows the user to choose whether to upgrade the original large model according to their actual needs, further enhancing the user experience.

[0063] In some embodiments of this application, the above-described large model upgrade method applied to vehicle infotainment systems further includes: in response to determining that a first result has been successfully obtained, releasing the original large model through the large model inference engine of the vehicle infotainment system.

[0064] In this implementation, if the vehicle-mounted system has successfully obtained the first result, it proves that the target large model can be used normally on the vehicle-mounted system and has generated inference results for the current user. At this time, the original large model will be released through the large model inference engine on the vehicle-mounted system, that is, the resources occupied by the original large model on the vehicle-mounted system will be released, thereby saving the storage resources of the vehicle-mounted system.

[0065] Figure 3This is a schematic diagram of the application process of the large-scale model upgrade method for in-vehicle systems according to this application, such as... Figure 3 As shown, the large-scale model upgrade method 300 applied to the vehicle system includes a step 301 executed in the cloud and a part 302 executed on the vehicle system.

[0066] Of which, the portion 310 executed in the cloud includes:

[0067] 1) Training a large model involves training a new large model based on the original large model, and then quantizing the trained large model to generate a quantized large model (the target large model). The target large model is then packaged, primarily by converting its format to enable it to run on the vehicle's infotainment chip platform. The packaging tool is typically provided by the chip platform.

[0068] 2) Save the large model. Store the packaged large model in the model repository. Specifically, store the compressed package of the large model file and the version information of the large model in the model repository.

[0069] 3) Large Model Upgrade Management Service. Specifically, the large model update management service creates an upgrade plan for the vehicle-side large language model and formulates an upgrade strategy. The upgrade strategy is typically divided into two parts: small-batch upgrade verification and large-batch upgrade push.

[0070] Small-batch upgrade verification involves selecting a subset of test vehicles from all vehicles for upgrade verification, primarily to confirm whether the new model can function correctly on the vehicle's infotainment system. Once the small-batch upgrade verification is complete, large-batch upgrade rollout begins.

[0071] The part 302 executed on the vehicle's infotainment system includes:

[0072] 4) Large model upgrade task: After the vehicle terminal receives the upgrade task pushed by the cloud large model upgrade management service, the user selects to confirm the upgrade, and the vehicle terminal starts to download the large language model upgrade package from the cloud model repository and saves it to the local repository.

[0073] In addition, the use of the large model will also be performed on the vehicle's infotainment system, meaning that the parts executed on the vehicle's infotainment system also include:

[0074] 1) Receive user input information.

[0075] 2) Combine the user's input with the locally stored prompt to obtain the prompt word.

[0076] 3) Input the combined content (i.e. prompt words) into the large model inference engine so that the large model inference engine loads the large model stored in the local warehouse of the vehicle terminal and starts to perform inference generation. After the inference is completed, the generated content is returned to the application based on the large language model (i.e. large model application).

[0077] 4) Large model applications execute related services based on the inference content, such as calling the interfaces of other applications or using TTS to convert text into speech for playback.

[0078] This application also provides a large-model upgrade device for vehicle systems, such as... Figure 4 As shown, the device includes an acquisition and operation module 401 and a loading and acquisition module 402. The acquisition and operation module 401 is configured to acquire and run a target large model in response to receiving a first instruction confirming the upgrade of the original large model during the reasoning process of the original large model loaded on the vehicle-mounted system based on prompt words. The loading and acquisition module 402 is configured to, in response to determining that the target large model has been successfully run and that intermediate results have been obtained through reasoning using the original large model, load the target large model to continue reasoning based on the prompt words and intermediate results, and obtain a first result.

[0079] In some implementations, the acquisition and operation module 401 is further configured to: load the original large model based on prompt words through the vehicle's large model inference engine; and in response to receiving the first instruction, load the original large model based on prompt words through the large model inference engine while running the target large model through the large model inference engine.

[0080] In some implementations, the loading and acquisition module 402 is further configured to: load the target large model through the vehicle's large model inference engine, continue inference based on prompt words and intermediate results, and obtain the first result.

[0081] In some implementations, the large model upgrade device 400 further includes an engine acquisition module configured to, in response to obtaining a prompt word through a large model application installed in the vehicle infotainment system, input the prompt word into the large model inference engine of the vehicle infotainment system via the application programming interface (API) of the large model application.

[0082] In some embodiments, the large model upgrade device 400 further includes: a download module configured to, in response to receiving a large model upgrade instruction including upgrade instruction information from the vehicle system, download the target large model and store it in the large model library at the moment indicated by the upgrade instruction information to acquire the target large model; wherein, the acquisition and running module 401 is further configured to, in response to receiving a first instruction, acquire and run the target large model from the large model library through the large model inference engine of the vehicle system.

[0083] In some embodiments, the large model upgrade device 400 further includes: a display module configured to display information requesting an upgrade of the original large model at the time indicated by the upgrade instruction information; and a confirmation module configured to obtain a first instruction based on a confirmation instruction for the information requesting an upgrade of the original large model.

[0084] In some implementations, the large model upgrade device 400 further includes a release module configured to release the original large model via the vehicle's large model inference engine in response to determining that a first result has been successfully obtained.

[0085] Embodiments of this application also provide an electronic device, such as... Figure 5 As shown, the electronic device 500 may include: at least one processor and a memory, the memory being communicatively connected to the at least one processor and storing instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the large-scale model upgrade method for vehicle systems mentioned in the above embodiments.

[0086] One embodiment of this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the large-scale model upgrade method for vehicle systems mentioned in the above embodiments.

[0087] Figure 5 This is a schematic block diagram of an electronic device 500 according to some embodiments of this application. For example... Figure 5 As shown, the electronic device 500 includes a processor 501, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 502 or a computer program loaded from a memory 508 into a random access memory (RAM) 503. The RAM 503 may also store various programs and data required for the operation of the electronic device 500. The processor 501, ROM 502, and RAM 503 are interconnected via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0088] Multiple components in electronic device 500 are connected to I / O interface 505, including: input unit 506, such as buttons or touchscreens in a vehicle infotainment system; output unit 507, connected to various types of displays, speakers, etc., to output various forms of signals; memory 508, including any medium for storing computer-executable programs; and communication unit 509, such as a network interface card (NIC), modem, or wireless transceiver. Communication unit 509 allows electronic device 500 to exchange information / data with other devices via a local area network (LAN) or other wireless communication networks.

[0089] Processor 501 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 501 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 501 performs the various methods and processes described above, such as the large-scale model upgrade method for in-vehicle systems mentioned in the above embodiments. For example, in some embodiments, the large-scale model upgrade method for in-vehicle systems mentioned in the above embodiments can be implemented as a computer software program, which is tangibly contained in a computer-readable storage medium, such as memory 508. In some embodiments, part or all of the computer program can be loaded and / or installed on electronic device 500 via ROM 502 and / or communication unit 509. When the computer program is loaded into RAM 503 and executed by processor 501, one or more steps of the large-scale model upgrade method for in-vehicle systems mentioned in the above embodiments can be performed. Alternatively, in other embodiments, the processor 501 may be configured by any other suitable means (e.g., by means of firmware) to perform the large-scale model upgrade method for vehicle systems mentioned in the above embodiments.

[0090] Various aspects of this application have been described herein with reference to flowchart illustrations and / or timing diagrams of methods, apparatus (systems), and computer program products according to exemplary embodiments of this application. It should be understood that each step of the flowchart illustrations and / or timing diagrams, as well as combinations of steps in the flowchart illustrations and / or timing diagrams, can be implemented by computer-readable program instructions.

[0091] These computer-readable program instructions can be provided to a processor, general-purpose computer, special-purpose computer, or other programmable data processing unit in an electronic device to produce a machine such that, when executed by the processing unit of the computer or other programmable data processing device, they create means for implementing the functions / steps specified in one or more steps of a flowchart and / or timing diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing device, and / or other device to operate in a particular manner. Thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / steps specified in one or more steps of a flowchart and / or timing diagram.

[0092] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to perform the functions / steps specified in one or more steps of a flowchart and / or timing diagram.

[0093] The flowcharts and timing diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of devices, methods, and computer program products according to various embodiments of this application. In this regard, each step in a flowchart or timing diagram may represent a module, segment, or part of an instruction that contains one or more executable instructions for implementing a specified logical function. In some alternative embodiments, the functions indicated in the steps may occur in a different order than those indicated in the drawings. For example, two consecutive steps may actually be performed substantially in parallel, and they may sometimes be performed in reverse order, depending on the functions involved. It should also be noted that each step in a timing diagram and / or flowchart, and combinations of steps in timing diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0094] The above description is merely an illustration of the embodiments of this application and the technical principles employed. Those skilled in the art should understand that the scope of protection involved in this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the technical concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A method for upgrading a large-scale vehicle infotainment system, comprising: During the process of loading the original large model on the vehicle system and performing reasoning based on prompt words, in response to the first instruction to confirm the upgrade of the original large model, the target large model is obtained and run. In response to determining that the target large model has been successfully run and that intermediate results have been obtained by loading the original large model, the target large model is loaded to continue reasoning based on the prompt words and the intermediate results to obtain the first result.

2. The method according to claim 1, wherein, During the process of loading the original large model on the vehicle system and performing reasoning based on prompt words, in response to receiving a first instruction confirming the upgrade of the original large model, the target large model is acquired and run, including: The vehicle's large-model inference engine loads the original large-model and performs inference based on the prompt words. In response to receiving the first instruction, the target large model is run simultaneously by loading the original large model through the large model inference engine and performing inference based on the prompt words.

3. The method according to claim 1 or 2, wherein, The loading of the target large model continues reasoning based on the prompt words and the intermediate results to obtain a first result, including: The target large model is loaded through the vehicle's large model inference engine, and inference is continued based on the prompt words and the intermediate results to obtain the first result.

4. The method according to claim 1 or 2, wherein, The method further includes: In response to obtaining the prompt word through a large model application installed in the vehicle system, the prompt word is input to the large model inference engine of the vehicle system through the application programming interface (API) of the large model application.

5. The method according to claim 1 or 2, wherein, The method further includes: In response to receiving a large model upgrade instruction including upgrade instruction information of the vehicle system, at the time indicated by the upgrade instruction information to obtain the target large model, the target large model is downloaded and stored in the large model library; The step of obtaining and running the target large model in response to receiving a first instruction confirming the upgrade of the original large model includes: In response to receiving the first instruction, the target large model is retrieved from the large model library and run through the large model inference engine of the vehicle system.

6. The method according to claim 5, wherein, The method further includes: At the moment indicated by the upgrade instruction information when the original large model is to be upgraded, information requesting the upgrade of the original large model is displayed; The first instruction is obtained based on the confirmation instruction regarding the information of the request to upgrade the original large model.

7. The method according to claim 1 or 2, wherein, The method further includes: In response to determining that the first result has been successfully obtained, the original large model is released through the large model inference engine of the vehicle system.

8. A large-scale model upgrade device, applied to vehicle infotainment systems, comprising: The acquisition and execution module is configured to, in response to a first instruction confirming the upgrade of the original large model on the vehicle system, acquire and run the target large model during the process of reasoning based on prompt words using the original large model loaded on the vehicle system; The loading module is configured to, in response to determining that the target large model has been successfully run and that intermediate results have been obtained by loading the original large model, load the target large model to continue reasoning based on the prompt words and the intermediate results, and obtain a first result.

9. An electronic device, comprising: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-7.

10. A non-transitory computer-readable storage medium storing computer instructions for causing the computer to perform the method of any one of claims 1-7.

11. A computer program product comprising a computer program that, when executed by a processor, implements the steps of the method according to any one of claims 1-7.

Images