Large model updating method and device, storage medium, electronic equipment and vehicle

Abstract

The invention discloses a large model updating method and device, a storage medium, electronic equipment and a vehicle, and relates to the technical field of artificial intelligence processing, and the method comprises the steps: downloading a target offline large model to a data partition of a vehicle memory when an offline large model in a preset partition of the vehicle memory needs to be updated, the off-line large model is an initial large model preset in the preset partition before the vehicle leaves a factory or is obtained by updating based on the initial large model, and the target off-line large model is used for updating the off-line large model in the preset partition. Compared with the prior art, the vehicle task can be executed on the basis of the target off-line large model under the off-line condition, so that the vehicle task can be executed under the condition that the vehicle leaves a factory and is not networked, and the user experience is improved.

Description

Technical Field

[0001] The present disclosure relates to the field of artificial intelligence processing technology, and in particular, to a large model update method, device, storage medium, electronic device, and vehicle. Background Art

[0002] The large model on the vehicle end refers to a machine learning or deep learning model deployed on the computing platform inside the vehicle. These models are used to perform various tasks, such as autonomous driving, advanced driver assistance systems (ADAS), in-vehicle experience optimization, etc. Summary of the Invention

[0003] In view of this, the present disclosure provides a large model update method, device, storage medium, electronic device, and vehicle to solve related technical problems.

[0004] According to the first aspect of the present disclosure, a large model update method is provided, including:

[0005] When the offline large model in the preset partition of the vehicle memory needs to be updated, download the target offline large model to the data partition of the vehicle memory. The offline large model is the initial large model preset in the preset partition before the vehicle leaves the factory or the one updated based on the initial large model, and the target offline large model is used to update the offline large model in the preset partition.

[0006] Optionally, after downloading the target offline large model to the data partition of the vehicle memory, the method further includes:

[0007] When the first condition is satisfied, update the offline large model in the preset partition to the target offline large model;

[0008] The first condition includes at least one of the following:

[0009] The size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition;

[0010] The integrity check of the target offline large model passes.

[0011] Optionally, updating the offline large model in the preset partition to the target offline large model includes:

[0012] Delete the offline large model to be updated in the preset partition, and copy the target offline large model from the data partition to the preset partition.

[0013] Optionally, after copying the target offline large model from the data partition to the preset partition, the method further includes:

[0014] Delete the target offline large model in the data partition.

[0015] Optionally, after updating the offline large model in the preset partition to the target offline large model, the method further includes:

[0016] Mark the target offline large model in the preset partition with a target identifier, where the target identifier indicates that the target offline large model is the most recently updated offline large model.

[0017] Optionally, the method further includes:

[0018] In the case where the size of the target offline large model is greater than the storage space threshold, mark the target offline large model in the data partition with a target identifier.

[0019] Optionally, the method further includes:

[0020] In the case where the size of the target offline large model is greater than the storage space threshold, mark the target offline large model in the data partition with a target identifier.

[0021] Optionally, based on the offline large model marked with the target identifier, perform vehicle tasks.

[0022] Optionally, the method further includes:

[0023] In the case where the size of the target offline large model is greater than the storage space threshold, based on the target offline large model in the data partition, perform vehicle tasks and mark the target offline large model in the data partition with the target identifier.

[0024] Optionally, the method further includes:

[0025] In the case where the integrity check fails, delete the target offline large model in the data partition.

[0026] According to a second aspect of the present disclosure, there is provided a large model update device, including:

[0027] A download module, configured to download a target offline large model to a data partition of the vehicle memory when the offline large model in a preset partition of the vehicle memory needs to be updated, where the offline large model is an initial large model preset in the preset partition before the vehicle leaves the factory or an updated model based on the initial large model, and the target offline large model is used to update the offline large model in the preset partition.

[0028] According to a third aspect of the present disclosure, there is provided a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the large model updating method described in the first aspect.

[0029] According to a fourth aspect of the present disclosure, there is provided an electronic device, including a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein when the processor executes the computer program, the large model updating method described in the first aspect is implemented.

[0030] In a fifth aspect, the present disclosure provides a vehicle, including the large model updating device described in the second aspect, or the computer-readable storage medium described in the third aspect, or the electronic device described in the fourth aspect.

[0031] By means of the above technical solutions, a large model updating method, device, storage medium, electronic device, and vehicle provided by the present disclosure, compared with the related art, in the case where it is detected that the offline large model in the preset partition of the vehicle memory needs to be updated, the target offline large model is downloaded to the data partition of the vehicle memory, and when the size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition, the offline large model in the preset partition is updated to the target offline large model. The offline large model in the preset partition can be updated, solving the situation that it is difficult to temporarily store or decompress the updated target offline large model due to limited preset space, thereby causing difficulties in updating the preset partition. Moreover, an initial large model can be configured in the preset partition before the vehicle leaves the factory, so that vehicle tasks can be executed even when the vehicle leaves the factory without being connected to the network, improving the user experience.

[0032] The above description is only an overview of the technical solutions of the present disclosure. In order to be able to understand the technical means of the present disclosure more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features, and advantages of the present disclosure more obvious and understandable, the following specifically illustrates the specific embodiments of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] The accompanying drawings herein are incorporated into the specification and form a part of the specification, showing embodiments consistent with the present disclosure and used together with the specification to explain the principles of the present disclosure.

[0034] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, for those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0035] Figure 1 It is a schematic flowchart of a large model updating method provided for an embodiment of the present disclosure;

[0036] Figure 2 Schematic flowchart of a large model update method provided by an embodiment of the present disclosure;

[0037] Figure 3 Schematic flowchart of an example provided by an embodiment of the present disclosure;

[0038] [[ID=,10]] Figure 4 Schematic structural diagram of a large model update device provided by an embodiment of the present disclosure

[0039] Figure 5 Shows a schematic structural diagram of a vehicle provided by an embodiment of the present application. Detailed implementation manners

[0040] The following describes exemplary embodiments of the present disclosure with reference to the accompanying drawings. Various details of the embodiments of the present disclosure are included to facilitate understanding and should be considered merely exemplary. Therefore, those of ordinary skill in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted below. It should be noted that, without conflict, the embodiments of the present disclosure and the features in the embodiments can be combined with each other.

[0041] The following describes the large model update method, device, storage medium, electronic device, and vehicle of the embodiments of the present disclosure with reference to the accompanying drawings.

[0042] To solve the problem in the related art that a vehicle needs to be connected to the network to use the functions corresponding to the large model, which affects the user experience, this embodiment provides a large model update method, as Figure 1 shown, the method includes:

[0043] Step 101: When the offline large model in the preset partition of the vehicle memory needs to be updated, download the target offline large model to the data partition of the vehicle memory.

[0044] Among them, the offline large model is the initial large model preset in the preset partition before the vehicle leaves the factory or the one updated based on the initial large model, and the target offline large model is used to update the offline large model in the preset partition.

[0045] In the embodiments of the present application, the preset partition can be a storage partition preconfigured in the vehicle system for storing the initial large model before the vehicle leaves the factory. Specifically, by presetting the initial large model in the preset partition, conflicts with other software components can be avoided, and it can be ensured that the core driving function can still work properly even when the operating system or other applications fail. The preset initial large model can also be directly loaded from a fast storage medium (such as an SSD or eMMC), reducing the startup time and response latency. The independent partition helps to manage and update the model more conveniently without affecting other parts of the system.

[0046] In some examples, the initial large model is preset in the preset partition before the vehicle leaves the factory. The offline large model can be the initial large model or an updated model based on the initial large model. Specifically, before the vehicle leaves the factory, the offline large model can be preset in the preset storage partition. These models are crucial for implementing advanced driver assistance systems (ADAS) or autonomous driving functions. They can perform key tasks such as environmental perception, path planning, and decision-making without a network connection.

[0047] Exemplarily, if the initial large model preset in the preset partition before vehicle A leaves the factory is large model A, then the offline large model can be large model A or large model B obtained by updating large model A at least once; correspondingly, if the initial large model preset in the preset partition before vehicle B leaves the factory is large model B, then the offline large model can be large model B or large model C obtained by updating large model B at least once.

[0048] In the embodiments of the present application, deploying the offline large model on the vehicle side directly affects the vehicle's functional performance (such as autonomous driving, driver assistance systems, etc.). Timely updates can improve performance, fix vulnerabilities, or adapt to new environmental requirements. Specifically, each model has a version number (such as v1.0.0). It can be determined whether an update is needed by comparing the current model version with the latest version in the cloud; it can also be determined whether the model needs to be optimized or updated by monitoring the actual running performance of the model (such as accuracy, latency, error rate, etc.); automatic updates can also be checked at preset time intervals (such as monthly or quarterly) to ensure that the model is always in a relatively new state; it can also be determined whether the model needs to be improved based on the user's actual usage experience, and no more examples will be given here.

[0049] In some examples, the vehicle end downloads the latest large model from the cloud to the local usually through over-the-air (OTA) updates to ensure that the vehicle can obtain the latest algorithm improvements, function expansions, or performance optimizations.

[0050] Exemplarily, the specific process for the vehicle end to download the latest large model from the cloud may include, but is not limited to: (1) Checking for updates: The vehicle sends a request to the cloud through the OTA mechanism to query whether there is a new model version. If there is a new version, it receives the meta-information returned by the cloud (such as version number, file size, checksum, etc.). (2) Network connection management: Ensure that the vehicle is in a stable network environment (such as Wi-Fi or cellular network). If the network conditions are poor, it can choose to delay the download or prompt the user to select a suitable time. (3) Downloading the model file: The vehicle end downloads the model file from the cloud through the HTTP / HTTPS protocol. The breakpoint resume function is supported to avoid download failure caused by network interruption.

[0051] In the embodiments of the present application, the preset partition generally has size limitations, so it cannot be used without limit. That is, the model package downloaded and updated, which is the target offline large model in the embodiments of the present application, is not directly stored and decompressed into the preset partition. That is, it is necessary to compare the size of the target offline large model with the storage space threshold corresponding to the preset partition. Only when it is determined that the size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition, can the target offline large model be updated to the preset partition.

[0052] In some examples, the target offline large model can be the large model of the new version in the cloud or the server that needs to be downloaded to the vehicle. For example, if the offline large model in the preset partition of the current vehicle A is large model A and the vehicle needs to update large model A, it can download the latest version of large model B from the cloud to the data partition of the vehicle. At this time, large model B is the target offline large model in the embodiments of the present application; similarly, if the offline large model in the preset partition of the current vehicle B is large model B and the vehicle needs to update large model B, it can download the latest version of large model C from the cloud to the data partition of the vehicle. At this time, large model C is the target offline large model in the embodiments of the present application.

[0053] It should be noted that after the target offline large model is updated to the preset partition, the vehicle can execute vehicle tasks by calling the target offline large model in the preset partition when the vehicle is not connected to the network. Specifically, it may include, but is not limited to, the advanced driver assistance system (ADAS) or the autonomous driving function, etc.

[0054] For this embodiment, the target offline large model downloaded to the preset partition can be used to update the offline large model in the preset partition under certain conditions; exemplarily, if the downloaded target offline large model B does not meet the conditions, then the target offline large model B cannot be used to update the offline large model in the preset partition; on the contrary, if the downloaded target offline large model C meets the conditions, then the target offline large model C can be used to update the offline large model in the preset partition.

[0055] Compared with the related art, in this embodiment, when it is detected that the offline large model in the preset partition of the vehicle memory needs to be updated, the target offline large model is downloaded to the data partition of the vehicle memory, and when the size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition, the offline large model in the preset partition is updated to the target offline large model. The offline large model in the preset partition can be updated, which solves the problem that it is difficult to temporarily store or decompress the updated target offline large model due to limited preset space, thus making it difficult to update the preset partition. Moreover, an initial large model can be configured in the preset partition before the vehicle leaves the factory, so that vehicle tasks can be executed even when the vehicle is not connected to the network after leaving the factory, improving the user experience.

[0056] Further, as a refinement and extension of the above embodiment, in order to fully illustrate the specific implementation process of the method in this embodiment, this embodiment provides the specific method as shown in Figure 2 The following steps are included:

[0057] Step 201: When the offline large model in the preset partition of the vehicle memory needs to be updated, download the target offline large model to the data partition of the vehicle memory.

[0058] Among them, the offline large model is the initial large model preset in the preset partition before the vehicle leaves the factory or the one obtained by updating based on the initial large model.

[0059] In the embodiment of the present application, to implement the offline large model capability on the vehicle and have it available even when the vehicle leaves the factory without being connected to the network, the offline large model will be preset in the preset partition of the vehicle and carried when leaving the factory. Therefore, the currently detected offline large model can be the initial large model preset in the preset partition before the vehicle leaves the factory. Since the large model stored in the preset partition will not be deleted due to a factory reset, the currently detected offline large model can also be the large model stored in the preset partition after being updated based on the initial large model.

[0060] In some examples, when the offline large model needs to be updated, the model needs to be downloaded to the data partition of the vehicle first. Among them, the data partition can separate the core model from the operating system, user data, etc., to avoid mutual interference; protect the key model from being tampered with or damaged; optimize the storage access speed and reduce the loading delay; support convenient updates, rollbacks, and expansions.

[0061] Exemplarily, if the vehicle terminal detects that there is a latest version of the offline large model in the cloud, that is, the target offline large model in the embodiment of the present application, the target offline large model can be downloaded to the local data partition.

[0062] Step 202: Update the offline large model in the preset partition to the target offline large model when the first condition is met.

[0063] Among them, the first condition includes at least one of the following:

[0064] The size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition;

[0065] The integrity check of the target offline large model passes.

[0066] It should be noted that performing an integrity check on the offline large model is an important step to ensure that the model file has not been tampered with or damaged during storage, transmission, and loading. Especially in the vehicle environment, since it involves key functions such as autonomous driving and driver assistance systems, the integrity and reliability of the model are crucial. Performing an integrity check can ensure that the model file has not been maliciously modified, can also identify file damage caused by reasons such as network interruption and storage failure, and confirm that the model file comes from a trusted cloud server.

[0067] Exemplarily, the integrity check methods can include but are not limited to: 1. Hash check: Use a hash algorithm (such as MD5, SHA-256) to calculate the hash value of the model file. Compare the calculated hash value with the standard hash value provided by the cloud. For example, use the Message Digest Algorithm 5 (MD5) for verification. Among them, MD5 is a widely used hash function that can map input data of any length to an output of a fixed length (128 bits, that is, 16 bytes), usually represented in the form of a 32-bit hexadecimal string. 2. Digital signature verification: Generate a digital signature for the model file using a private key in the cloud, and the vehicle end uses the corresponding public key to verify the validity of the signature. 3. Redundancy check (CRC check): Use the Cyclic Redundancy Check (CRC) algorithm to calculate the check code, and compare the locally calculated CRC value with the CRC value provided by the cloud.

[0068] In an embodiment of the present application, the embodiment of the present application may specifically include: when the offline large model in the preset partition of the vehicle memory needs to be updated, downloading the target offline large model to the data partition of the vehicle memory, performing an integrity check on the target offline large model, and deleting the target offline large model in the data partition if the integrity check of the target offline large model fails; when the integrity check of the target offline large model passes, determining the size of the target offline large model, and if it is determined that the size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition, deleting the offline large model that needs to be updated in the preset partition, copying the target offline large model from the data partition to the preset partition, deleting the target offline large model in the data partition, and marking the target offline large model in the preset partition with a target identifier; if it is determined that the size of the target offline large model is greater than the storage space threshold, marking the target offline large model in the data partition with a target identifier, and performing vehicle tasks based on the offline large model marked with the target identifier.

[0069] In some examples, when the first condition is that "the size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition", the method of this embodiment may include:

[0070] In some examples, when the size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition is satisfied, updating the offline large model in the preset partition to the target offline large model.

[0071] In some examples, when the size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition is satisfied, deleting the offline large model that needs to be updated in the preset partition, and copying the target offline large model from the data partition to the preset partition.

[0072] In some examples, when the size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition is satisfied, deleting the offline large model that needs to be updated in the preset partition, copying the target offline large model from the data partition to the preset partition, and deleting the target offline large model in the data partition.

[0073] In some examples, when the size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition is satisfied, updating the offline large model in the preset partition to the target offline large model, marking the target offline large model in the preset partition with a target identifier, and the target identifier indicates that the target offline large model is the most recently updated offline large model.

[0074] In some examples, when the size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition, the offline large model that needs to be updated in the preset partition is deleted, the target offline large model is copied from the data partition to the preset partition, and the target offline large model in the preset partition is marked with a target identifier, where the target identifier indicates that the target offline large model is the most recently updated offline large model.

[0075] In some examples, when the size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition, the offline large model that needs to be updated in the preset partition is deleted, the target offline large model is copied from the data partition to the preset partition, the target offline large model in the data partition is deleted, the target offline large model is copied from the data partition to the preset partition, and the target offline large model in the preset partition is marked with a target identifier, where the target identifier indicates that the target offline large model is the most recently updated offline large model.

[0076] In some examples, when the size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition, the offline large model in the preset partition is updated to the target offline large model; when the size of the target offline large model is greater than the storage space threshold, the target offline large model in the data partition is marked with a target identifier.

[0077] In some examples, when the size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition, the offline large model that needs to be updated in the preset partition is deleted, the target offline large model is copied from the data partition to the preset partition; when the size of the target offline large model is greater than the storage space threshold, the target offline large model in the data partition is marked with a target identifier.

[0078] In some examples, when the size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition, the offline large model that needs to be updated in the preset partition is deleted, the target offline large model is copied from the data partition to the preset partition, the target offline large model in the data partition is deleted; when the size of the target offline large model is greater than the storage space threshold, the target offline large model in the data partition is marked with a target identifier.

[0079] In some examples, when the size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition, the offline large model in the preset partition is updated to the target offline large model, the target offline large model in the preset partition is marked with a target identifier, where the target identifier indicates that the target offline large model is the most recently updated offline large model, and vehicle tasks are executed based on the offline large model marked with the target identifier.

[0080] In some examples, when the size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition, the offline large model that needs to be updated in the preset partition is deleted, the target offline large model is copied from the data partition to the preset partition, and the target offline large model in the preset partition is marked with a target identifier, where the target identifier indicates that the target offline large model is the most recently updated offline large model. Based on the offline large model marked with the target identifier, vehicle tasks are executed.

[0081] In some examples, when the size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition, the offline large model that needs to be updated in the preset partition is deleted, the target offline large model is copied from the data partition to the preset partition, the target offline large model in the data partition is deleted, the target offline large model is copied from the data partition to the preset partition, and the target offline large model in the preset partition is marked with a target identifier, where the target identifier indicates that the target offline large model is the most recently updated offline large model. Based on the offline large model marked with the target identifier, vehicle tasks are executed.

[0082] In some examples, when the size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition, the offline large model in the preset partition is updated to the target offline large model. When the size of the target offline large model is greater than the storage space threshold, the target offline large model in the data partition is marked with a target identifier. Based on the offline large model marked with the target identifier, vehicle tasks are executed.

[0083] In some examples, when the size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition, the offline large model that needs to be updated in the preset partition is deleted, the target offline large model is copied from the data partition to the preset partition. When the size of the target offline large model is greater than the storage space threshold, the target offline large model in the data partition is marked with a target identifier. Based on the offline large model marked with the target identifier, vehicle tasks are executed.

[0084] In some examples, when the size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition, the offline large model that needs to be updated in the preset partition is deleted, the target offline large model is copied from the data partition to the preset partition, the target offline large model in the data partition is deleted. When the size of the target offline large model is greater than the storage space threshold, the target offline large model in the data partition is marked with a target identifier. Based on the offline large model marked with the target identifier, vehicle tasks are executed.

[0085] In some examples, when the size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition, the offline large model in the preset partition is updated to the target offline large model. When the size of the target offline large model is greater than the storage space threshold, vehicle tasks are executed based on the target offline large model in the data partition, and the target offline large model in the data partition is marked with a target identifier.

[0086] In some examples, when the size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition, the offline large model to be updated in the preset partition is deleted, and the target offline large model is copied from the data partition to the preset partition. When the size of the target offline large model is greater than the storage space threshold, vehicle tasks are executed based on the target offline large model in the data partition, and the target offline large model in the data partition is marked with a target identifier.

[0087] In some examples, when the size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition, the offline large model to be updated in the preset partition is deleted, and the target offline large model is copied from the data partition to the preset partition. The target offline large model in the data partition is deleted. When the size of the target offline large model is greater than the storage space threshold, vehicle tasks are executed based on the target offline large model in the data partition, and the target offline large model in the data partition is marked with a target identifier.

[0088] In some examples, when the size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition, the offline large model in the preset partition is updated to the target offline large model. When the integrity check fails, the target offline large model in the data partition is deleted.

[0089] In some examples, when the size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition, the offline large model to be updated in the preset partition is deleted, and the target offline large model is copied from the data partition to the preset partition. When the integrity check fails, the target offline large model in the data partition is deleted.

[0090] In some examples, when the size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition, the offline large model to be updated in the preset partition is deleted, and the target offline large model is copied from the data partition to the preset partition. The target offline large model in the data partition is deleted. When the integrity check fails, the target offline large model in the data partition is deleted.

[0091] In some examples, when the first condition is that the integrity check of the target offline large model passes, the method of this embodiment may include:

[0092] In some examples, when the integrity check of the target offline large model passes, update the offline large model in the preset partition to the target offline large model.

[0093] In some examples, when the integrity check of the target offline large model passes, delete the offline large model to be updated in the preset partition, and copy the target offline large model from the data partition to the preset partition.

[0094] In some examples, when the integrity check of the target offline large model passes, delete the offline large model to be updated in the preset partition, copy the target offline large model from the data partition to the preset partition, and delete the target offline large model in the data partition.

[0095] In some examples, when the integrity check of the target offline large model passes, update the offline large model in the preset partition to the target offline large model, and mark the target offline large model in the preset partition with a target identifier, where the target identifier indicates that the target offline large model is the most recently updated offline large model.

[0096] In some examples, when the integrity check of the target offline large model passes, delete the offline large model to be updated in the preset partition, copy the target offline large model from the data partition to the preset partition, and mark the target offline large model in the preset partition with a target identifier, where the target identifier indicates that the target offline large model is the most recently updated offline large model.

[0097] In some examples, when the integrity check of the target offline large model passes, delete the offline large model to be updated in the preset partition, copy the target offline large model from the data partition to the preset partition, delete the target offline large model in the data partition, copy the target offline large model from the data partition to the preset partition, and mark the target offline large model in the preset partition with a target identifier, where the target identifier indicates that the target offline large model is the most recently updated offline large model.

[0098] In some examples, when the integrity check of the target offline large model passes, update the offline large model in the preset partition to the target offline large model, and when the size of the target offline large model is greater than the storage space threshold, mark the target offline large model in the data partition with a target identifier.

[0099] In some examples, when the integrity check of the target offline large model passes, delete the offline large model to be updated in the preset partition, copy the target offline large model from the data partition to the preset partition, and when the size of the target offline large model is greater than the storage space threshold, mark the target offline large model in the data partition with a target identifier.

[0100] In some examples, when the integrity check of the target offline large model passes, delete the offline large model to be updated in the preset partition, copy the target offline large model from the data partition to the preset partition, delete the target offline large model in the data partition, and when the size of the target offline large model is greater than the storage space threshold, mark the target offline large model in the data partition with a target identifier.

[0101] In some examples, when the integrity check of the target offline large model passes, update the offline large model in the preset partition to the target offline large model, mark the target offline large model in the preset partition with a target identifier, where the target identifier indicates that the target offline large model is the most recently updated offline large model, and based on the offline large model marked with the target identifier, perform vehicle tasks.

[0102] In some examples, when the integrity check of the target offline large model passes, delete the offline large model to be updated in the preset partition, copy the target offline large model from the data partition to the preset partition, mark the target offline large model in the preset partition with a target identifier, where the target identifier indicates that the target offline large model is the most recently updated offline large model, and based on the offline large model marked with the target identifier, perform vehicle tasks.

[0103] In some examples, when the integrity check of the target offline large model passes, delete the offline large model to be updated in the preset partition, copy the target offline large model from the data partition to the preset partition, delete the target offline large model in the data partition, copy the target offline large model from the data partition to the preset partition, mark the target offline large model in the preset partition with a target identifier, where the target identifier indicates that the target offline large model is the most recently updated offline large model, and based on the offline large model marked with the target identifier, perform vehicle tasks.

[0104] In some examples, when the integrity check of the target offline large model passes, update the offline large model in the preset partition to the target offline large model, and when the size of the target offline large model is greater than the storage space threshold, mark the target offline large model in the data partition with a target identifier, and based on the offline large model marked with the target identifier, perform vehicle tasks.

[0105] In some examples, when the integrity check of the target offline large model passes, delete the offline large model to be updated in the preset partition, copy the target offline large model from the data partition to the preset partition, and when the size of the target offline large model is greater than the storage space threshold, mark the target offline large model in the data partition with a target identifier, and perform vehicle tasks based on the offline large model marked with the target identifier.

[0106] In some examples, when the integrity check of the target offline large model passes, delete the offline large model to be updated in the preset partition, copy the target offline large model from the data partition to the preset partition, delete the target offline large model in the data partition, and when the size of the target offline large model is greater than the storage space threshold, mark the target offline large model in the data partition with a target identifier, and perform vehicle tasks based on the offline large model marked with the target identifier.

[0107] In some examples, when the integrity check of the target offline large model passes, update the offline large model in the preset partition to the target offline large model, and when the size of the target offline large model is greater than the storage space threshold, perform vehicle tasks based on the target offline large model in the data partition, and mark the target offline large model in the data partition with a target identifier.

[0108] In some examples, when the integrity check of the target offline large model passes, delete the offline large model to be updated in the preset partition, copy the target offline large model from the data partition to the preset partition, and when the size of the target offline large model is greater than the storage space threshold, perform vehicle tasks based on the target offline large model in the data partition, and mark the target offline large model in the data partition with a target identifier.

[0109] In some examples, when the integrity check of the target offline large model passes, delete the offline large model to be updated in the preset partition, copy the target offline large model from the data partition to the preset partition, delete the target offline large model in the data partition, and when the size of the target offline large model is greater than the storage space threshold, perform vehicle tasks based on the target offline large model in the data partition, and mark the target offline large model in the data partition with a target identifier.

[0110] In some examples, when the integrity check of the target offline large model passes, update the offline large model in the preset partition to the target offline large model, and when the integrity check fails, delete the target offline large model in the data partition.

[0111] In some examples, when the integrity check of the target offline large model passes, the offline large model that needs to be updated in the preset partition is deleted, and the target offline large model is copied from the data partition to the preset partition. When the integrity check fails, the target offline large model in the data partition is deleted.

[0112] In some examples, when the integrity check of the target offline large model passes, the offline large model that needs to be updated in the preset partition is deleted, the target offline large model is copied from the data partition to the preset partition, and the target offline large model in the data partition is deleted. When the integrity check fails, the target offline large model in the data partition is deleted.

[0113] In the embodiments of the present application, if the integrity check fails, the target offline large model downloaded to the data partition needs to be deleted. Exemplarily, Figure 3 as shown, it can be a schematic flow chart for verification based on MD5.

[0114] It should be noted that when it is determined that the verification fails, if the file is damaged, try to download the latest version from the cloud again; if the model in the current partition is unavailable, switch to the backup partition or the old version; notify the user of the problem details through the in-vehicle system interface or the mobile phone App.

[0115] For this embodiment, when it is determined that the size of the target offline large model is less than or equal to the storage space threshold, the target offline large model in the data partition can be copied, pasted into the preset partition, and the target offline large model in the data partition is deleted. That is, when it is determined that the target offline large model can be stored in the preset partition, the target offline large model in the data partition is copied to the preset partition for storage, and the target offline large model stored in the preset partition is marked with the target identifier, that is, the target offline large model stored in the preset partition is marked as the offline large model that is newly downloaded to the vehicle terminal.

[0116] Exemplarily, if the target offline large model is Model A, when it is determined that Model A can be stored in the preset partition, Model A in the data partition is copied to the preset partition for storage, and Model A stored in the preset partition is marked as the offline large model that is newly downloaded to the vehicle terminal.

[0117] As an alternative, if it is determined that the size of the target offline large model is greater than the storage space threshold, that is, it is determined that the target offline large model cannot be stored in the preset partition, the target offline large model in the data partition is marked with the target identifier.

[0118] Exemplarily, if the target offline large model is Model A, when it is determined that Model A cannot be stored in the preset partition, Model A in the data partition is marked as the offline large model that is newly downloaded to the vehicle terminal.

[0119] Exemplarily, if the target offline large model is Model A, when it is determined that Model A can be stored in the preset partition, copy Model A in the data partition to the preset partition for storage, and mark the Model A stored in the preset partition as the offline large model that has been newly downloaded to the vehicle terminal. That is, when the vehicle is offline, call Model A stored in the preset partition to execute vehicle tasks; on the contrary, when it is determined that Model A cannot be stored in the preset partition (for example, the size of Model A is greater than the storage space threshold of the preset partition), mark Model A in the data partition as the offline large model that has been newly downloaded to the vehicle terminal. That is, when the vehicle is offline, call Model A stored in the data partition to execute vehicle tasks.

[0120] It should be noted that the offline large models stored in the data partition will be deleted due to factory reset, while the offline large models stored in the preset partition will not be deleted due to factory reset; Exemplarily, if it is determined that Model A cannot be stored in the preset partition, mark Model A in the data partition as the offline large model that has been newly downloaded to the vehicle terminal, and the offline large model stored in the preset partition is Model B that was updated to the preset partition last time. If the vehicle is factory reset, Model A in the data partition will be deleted, and the offline large model used when the vehicle is offline after leaving the factory and not connected to the network is Offline Large Model B.

[0121] It should be noted that with the development of large models, since the large models on the terminal side are also relatively large (in the GB level), it is impossible to package them with the Read-Only Memory (ROM) for release together (a too large ROM will affect the OTA success rate). Therefore, it is necessary to build the large models on the terminal side into the preset partition (pre-installed at the factory to provide large model capabilities). Generally, the preset partition has size limitations, so it cannot be used without limit. That is, the downloaded model package is not directly stored and decompressed into the preset partition, but downloaded to the data partition (which has a larger space). After verification, it is then synchronized to the preset partition. If the preset partition has reached its upper limit, the latest model is placed in the data partition to avoid affecting the preset partition.

[0122] It should be noted that the large model update method provided in the embodiments of the present application can also be extended to other fields such as map library update, which is not specifically limited herein.

[0123] Compared with the related art, in this embodiment, when it is detected that the offline large model in the preset partition of the vehicle memory needs to be updated, the target offline large model is downloaded to the data partition of the vehicle memory, and when the size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition, the offline large model in the preset partition is updated to the target offline large model. The offline large model in the preset partition can be updated, solving the problem that it is difficult to temporarily store or decompress the updated target offline large model due to limited preset space, which in turn causes difficulties in updating the preset partition. Moreover, an initial large model can be configured in the preset partition before the vehicle leaves the factory, so that vehicle tasks can be executed even when the vehicle leaves the factory without being connected to the network, improving the user experience.

[0124] Based on the above Figure 1 specific implementation of the method shown, this embodiment provides a large model update device, as Figure 4 shown, the device includes: a download module 31;

[0125] The download module 31 is configured to download the target offline large model to the data partition of the vehicle memory when the offline large model in the preset partition of the vehicle memory needs to be updated. The offline large model is the initial large model preset in the preset partition before the vehicle leaves the factory or is updated based on the initial large model. The target offline large model is used to update the offline large model in the preset partition.

[0126] In some examples of this embodiment, the download module 31 is further configured to update the offline large model in the preset partition to the target offline large model when the first condition is satisfied; the first condition includes at least one of the following: the size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition; the integrity check of the target offline large model passes.

[0127] In some examples of this embodiment, the download module 31 is specifically configured to delete the offline large model that needs to be updated in the preset partition, and copy the target offline large model from the data partition to the preset partition.

[0128] In some examples of this embodiment, the download module 31 is specifically further configured to delete the target offline large model in the data partition.

[0129] In some examples of this embodiment, the download module 31 is specifically further configured to mark the target offline large model in the preset partition with a target identifier, and the target identifier indicates that the target offline large model is the offline large model updated most recently.

[0130] In some examples of this embodiment, the download module 31 is further specifically configured to, when the size of the target offline large model is greater than the storage space threshold, mark the target offline large model in the data partition with a target identifier.

[0131] In some examples of this embodiment, the download module 31 is further specifically configured to execute vehicle tasks based on the offline large model marked with the target identifier.

[0132] In some examples of this embodiment, the download module 31 is further specifically configured to, when the size of the target offline large model is greater than the storage space threshold, execute vehicle tasks based on the target offline large model in the data partition, and mark the target offline large model in the data partition with the target identifier.

[0133] In some examples of this embodiment, the download module 31 is further specifically configured to delete the target offline large model in the data partition when the integrity check fails.

[0134] It should be noted that for other corresponding descriptions of each functional unit involved in the large model update device provided in this embodiment, reference can be made to Figure 1 the corresponding description in the method, which will not be elaborated here.

[0135] Based on the method as described above Figure 1 shown, correspondingly, the present disclosure further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the method as described above Figure 1 shown is implemented.

[0136] Based on such an understanding, the technical solution of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods in various implementation scenarios of the present disclosure.

[0137] Based on the method as described above Figure 1 shown, and Figure 4 the virtual device embodiment shown, for the purpose of achieving the above object, the embodiments of the present disclosure further provide an electronic device, such as intelligent terminals such as smartphones, tablets, drones, and intelligent robots. The device includes a storage medium and a processor; the storage medium is used to store a computer program; the processor is used to execute the computer program to implement the method as described above Figure 1 shown.

[0138] Optionally, the above-mentioned physical device may further include a user interface, a network interface, a camera, a Radio Frequency (RF) circuit, sensors, an audio circuit, a WI-FI module, etc. The user interface may include a display screen (Display), an input unit such as a keyboard (Keyboard), etc. Optionally, the user interface may further include a USB interface, a card reader interface, etc. The network interface may optionally include a standard wired interface, a wireless interface (such as a WI-FI interface), etc.

[0139] Those skilled in the art can understand that the above-mentioned physical device structure provided by the present disclosure does not constitute a limitation on the physical device, and may include more or fewer components, or combine certain components, or have different component arrangements.

[0140] The storage medium may further include an operating system and a network communication module. The operating system is a program for managing the hardware and software resources of the above-mentioned physical device, and supports the operation of information processing programs and other software and / or programs. The network communication module is used to implement communication between components inside the storage medium, as well as communication between other hardware and software in the information processing physical device.

[0141] Figure 5 is a block diagram of a vehicle shown according to an exemplary embodiment. For example, the vehicle 500 may be a hybrid vehicle, or a non-hybrid vehicle, an electric vehicle, a fuel cell vehicle, or other types of vehicles. The vehicle 500 may be an autonomous vehicle, a semi-autonomous vehicle, or a non-autonomous vehicle.

[0142] Referring to Figure 5 , the vehicle 500 may include various subsystems. For example, the infotainment system 501, the perception system 502, the decision control system 503, the drive system 504, and the computing platform 505. Among them, the vehicle 500 may further include more or fewer subsystems, and each subsystem may include multiple components. In addition, each subsystem and each component of the vehicle 500 may be interconnected by wired or wireless means.

[0143] In some embodiments, the infotainment system 501 may include a communication system, an entertainment system, a navigation system, etc.

[0144] The perception system 502 may include several sensors for sensing information about the environment around the vehicle 500. For example, the perception system 502 may include a global positioning system (the global positioning system may be a GPS system, or a Beidou system, or other positioning systems), an inertial measurement unit (IMU), lidar, millimeter-wave radar, ultrasonic radar, and a camera device.

[0145] The decision control system 503 may include a computing system, a vehicle controller, a steering system, an accelerator, and a braking system.

[0146] The drive system 504 may include components that provide powered movement for the vehicle 500. In one embodiment, the drive system 504 may include an engine, an energy source, a transmission system, and wheels. The engine may be one or a combination of an internal combustion engine, an electric motor, and an air compression engine. The engine is capable of converting the energy provided by the energy source into mechanical energy.

[0147] Some or all functions of the vehicle 500 are controlled by the computing platform 505. The computing platform 505 may include at least one processor 551 and a memory 552. The processor 551 may execute instructions 553 stored in the memory 552.

[0148] The processor 551 may be any conventional processor, such as a commercially available CPU. The processor may also include, for example, a Graphic Process Unit (GPU), a Field Programmable Gate Array (FPGA), a System on Chip (SOC), an Application Specific Integrated Circuit (ASIC), or a combination thereof.

[0149] The memory 552 may be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic memory, flash memory, a magnetic disk, or an optical disk.

[0150] In addition to the instructions 553, the memory 552 may also store data, such as road maps, route information, data on the position, direction, speed, etc. of the vehicle. The data stored in the memory 552 can be used by the computing platform 505.

[0151] In an embodiment of the present disclosure, the processor 551 may execute the instructions 553 to implement all or part of the steps of the large model update method provided by the present disclosure.

[0152] Through the description of the above embodiments, those skilled in the art can clearly understand that the present disclosure can be implemented by means of software plus a necessary general hardware platform, or can also be implemented by hardware. Compared with the related art, in this embodiment, when it is detected that the offline large model in the preset partition of the vehicle memory needs to be updated, the target offline large model is downloaded to the data partition of the vehicle memory, and when the size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition, the offline large model in the preset partition is updated to the target offline large model. The offline large model in the preset partition can be updated, which solves the problem that it is difficult to temporarily store or decompress the updated target offline large model due to limited preset space, thereby causing difficulties in updating the preset partition. Moreover, an initial large model can be configured in the preset partition before the vehicle leaves the factory, so that vehicle tasks can be executed even when the vehicle leaves the factory without being connected to the network, improving the user experience.

[0153] It should be noted that in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising one..." does not exclude the existence of additional identical elements in the process, method, article or device comprising the element.

[0154] It should be noted that in this article, all operations related to artificial intelligence large models are carried out under the authorization of the user and strictly comply with relevant laws and regulations such as privacy and security.

[0155] The above are only specific embodiments of the present disclosure, enabling those skilled in the art to understand or implement the present disclosure. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present disclosure. Therefore, the present disclosure will not be limited to these embodiments herein, but will be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A large model update method, characterized in that, Comprising: When the offline large model in the preset partition of the vehicle memory needs to be updated, download the target offline large model to the data partition of the vehicle memory. The offline large model is the initial large model preset in the preset partition before the vehicle leaves the factory or is updated based on the initial large model. The target offline large model is used to update the offline large model in the preset partition.

2. The method according to claim 1, wherein After downloading the target offline large model to the data partition of the vehicle memory, the method further comprises: When a first condition is satisfied, update the offline large model in the preset partition to the target offline large model; The first condition includes at least one of the following: The size of the target offline large model is less than or equal to the storage space threshold corresponding to the preset partition; The integrity check of the target offline large model passes.

3. The method according to claim 2, wherein The step of updating the offline large model in the preset partition to the target offline large model includes: Delete the offline large model to be updated in the preset partition, and copy the target offline large model from the data partition to the preset partition.

4. The method according to claim 3, wherein After copying the target offline large model from the data partition to the preset partition, the method further comprises: Delete the target offline large model in the data partition.

5. The method according to any one of claims 2 to 4, characterized in that, After updating the offline large model in the preset partition to the target offline large model, the method further comprises: Mark the target offline large model in the preset partition with a target identifier, where the target identifier indicates that the target offline large model is the most recently updated offline large model.

6. The method according to claim 2, characterized in that The method further comprises: When the size of the target offline large model is greater than the storage space threshold, mark the target offline large model in the data partition with a target identifier.

7. The method according to claim 5 or 6, characterized in that, The method further comprises: Execute vehicle tasks based on the offline large model marked with the target identifier.

8. The method according to claim 2, wherein The method further comprises: When the size of the target offline large model is greater than the storage space threshold, execute vehicle tasks based on the target offline large model in the data partition, and mark the target offline large model in the data partition with the target identifier.

9. The method according to claim 2, characterized in that The method further comprises: When the integrity check fails, delete the target offline large model in the data partition.

10. A large model update device, characterized in that, Comprising: A download module configured to download the target offline large model to the data partition of the vehicle memory when the offline large model in the preset partition of the vehicle memory needs to be updated. The offline large model is the initial large model preset in the preset partition before the vehicle leaves the factory or is updated based on the initial large model. The target offline large model is used to update the offline large model in the preset partition.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method according to any one of claims 1 to 9.

12. An electronic device, comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method according to any one of claims 1 to 9.

13. A vehicle, characterized in that, Comprising the device according to claim 10, or the computer-readable storage medium according to claim 11, or the electronic device according to claim 12.

Images