Model Deployment Update Processing Method and Its Device, Equipment, Medium, Product

Abstract

The present application discloses a method and apparatus, device, medium, and product for model deployment and update processing. The method includes: responding to a model update event acting on an online service of a target model, and obtaining a new neural network model and its corresponding model configuration information; pushing a new model storage instruction to a shared container service, and driving the shared container service to respond to the instruction to store the new neural network model corresponding to a model version number in a target service storage space in a target model pool according to the model configuration information; when the online service of the target model meets a preset new model deployment policy, pushing a shared storage path to the online service of the target model, and driving the online service of the target model to obtain the new neural network model from the shared container service according to the shared storage path for deployment. The present application realizes automatic management of the deployment and update of the new neural network model, saves the deployment time of the new model, improves the model hot update efficiency, and isolates and manages the model data to ensure its data security.

Description

Technical Field

[0001] This application relates to the field of neural network models, and in particular to a method for processing model deployment and update. In addition, it also relates to a corresponding device, equipment, non-volatile storage medium, and computer program product for this method. Background Art

[0002] In existing Internet platforms, there are various types of online services built based on neural network models. For example, in an online service for providing intelligent customer service, there is generally a model service corresponding to a neural network model for semantic reasoning. In an online service for providing product classification for users, there is generally a model service corresponding to a neural network model for product classification. Moreover, with the development of neural network technology, each platform will also update the neural network models that provide inference functions in its online services to improve the business processing efficiency of the online services built based on neural network models and optimize the user experience.

[0003] However, when developers in existing platforms deploy a new neural network model to the corresponding model online service, they often need to manually deploy the new neural network model to the corresponding model online service for model update after completing the development and training of the new neural network model. The manual method not only has a slow execution efficiency, but also is prone to model deployment failures due to operation errors during the manual deployment process, affecting the stability of the model online service.

[0004] Secondly, although some platforms also have corresponding model automated deployment systems, they only write the operation methods of manual deployment into corresponding execution scripts to achieve automated deployment, and cannot effectively manage the neural network models of a large number of model online services in the platform.

[0005] In view of the problems existing in the deployment of neural network models in existing online services, the applicant has made corresponding explorations to solve this problem. Summary of the Invention

[0006] The purpose of this application is to provide a method for processing model deployment and update to meet user needs. In addition, it also relates to a corresponding device, equipment, non-volatile storage medium, and computer program product for this method.

[0007] To achieve the purpose of this application, the following technical solutions are adopted:

[0008] A method for processing model deployment and update proposed to adapt to the purpose of this application includes the following steps:

[0009] In response to a model update event acting on the target model online service, obtain the new neural network model and its corresponding model configuration information included in the event, where the model configuration information includes a model pool identifier, a model service identifier, and a model version number;

[0010] Push a new model storage instruction to the shared container service, driving the shared container service to respond to the instruction and store the new neural network model corresponding to the model version number in the target service storage space corresponding to the service identifier in the target model pool corresponding to the model pool identifier according to the model configuration information;

[0011] When the target model online service meets the preset new model deployment policy, push the shared storage path corresponding to the new neural network model to the target model online service, driving the target model online service to obtain the new neural network model from the shared container service according to the shared storage path for deployment.

[0012] In a further embodiment, in the step of responding to a model update event acting on the target model online service and obtaining the new neural network model and its corresponding model configuration information included in the event, it includes:

[0013] Receive the new neural network model and its corresponding model configuration information pushed by the model development end through the model upload interface;

[0014] Correspondingly store the new neural network model and the model configuration information in the model cloud storage space.

[0015] In a further embodiment, in the step of driving the shared container service to respond to the instruction and store the new neural network model corresponding to the model version number in the target service storage space corresponding to the service identifier in the target model pool corresponding to the model pool identifier according to the model configuration information, it includes the following steps executed by the shared container service:

[0016] Respond to the new model storage instruction pushed by the model update service, and obtain the new neural network model and its model configuration information corresponding to the instruction from the model cloud storage space;

[0017] Obtain the model pool identifier and the model service identifier included in the model configuration information, and query the target service storage space corresponding to the service identifier in the target model pool corresponding to the model pool identifier;

[0018] Obtain the model version number included in the model configuration information, and correspondingly store the new neural network model and the model version number in the target service storage space.

[0019] In a further embodiment, the step of pushing the shared storage path corresponding to the new neural network model to the target model online service when the target model online service meets the preset new model deployment policy includes:

[0020] Obtain the model update time preset for the target model online service, and monitor whether the current time exceeds the model update time;

[0021] When it is monitored that the current time exceeds the model update time, push a model update instruction to the target model online service, and the model update instruction includes the shared storage path.

[0022] In a further embodiment, the step of driving the target model online service to obtain the new neural network model from the shared container service according to the shared storage path for deployment includes the following steps executed by the target model online service:

[0023] Respond to the model update instruction pushed by the model update service, and obtain the shared storage path included in the instruction;

[0024] According to the shared storage path, obtain the new neural network model from the target service storage space in the target model pool of the shared container service;

[0025] After completing the deployment of the new neural network model, take the current old neural network model offline and bring the new neural network model online to provide an inference function for the current model service.

[0026] In a further embodiment, in the step of pushing the shared storage path corresponding to the new neural network model to the target model online service, the shared storage path is pushed by the shared container service to the current model update service, and the shared storage path points to the new neural network model in the target service storage space of the target model pool.

[0027] A model deployment update processing device proposed for the purpose of adapting to the present application includes:

[0028] An update event response module, configured to respond to a model update event acting on the target model online service, and obtain the new neural network model and its corresponding model configuration information included in the event, where the model configuration information includes a model pool identifier, a model service identifier, and a model version number;

[0029] A model shared storage module, configured to push a new model storage instruction to the shared container service, and drive the shared container service to respond to the instruction to store the new neural network model corresponding to the model version number in the target service storage space corresponding to the service identifier in the target model pool corresponding to the model pool identifier according to the model configuration information;

[0030] A new model deployment module, which is used to push the shared storage path corresponding to the new neural network model to the target model online service when the target model online service meets the preset new model deployment policy, and drive the target model online service to obtain the new neural network model from the shared container service according to the shared storage path for deployment.

[0031] In a further embodiment, the update event response module includes:

[0032] A model acquisition sub-module, which is used to receive the new neural network model and its corresponding model configuration information pushed by the model development end through the model upload interface;

[0033] A model cloud storage sub-module, which is used to store the new neural network model and the model configuration information correspondingly in the model cloud storage space.

[0034] In a further embodiment, the model shared storage module includes:

[0035] A storage instruction response sub-module, which is used to respond to the new model storage instruction pushed by the model update service, and obtain the new neural network model and its model configuration information corresponding to the instruction from the model cloud storage space;

[0036] A storage space query sub-module, which is used to obtain the model pool identifier and the model service identifier included in the model configuration information, and query the target service storage space corresponding to the service identifier in the target model pool corresponding to the model pool identifier;

[0037] A model storage sub-module, which is used to obtain the model version number included in the model configuration information, and store the new neural network model and the model version number correspondingly in the target service storage space.

[0038] In a further embodiment, the new model deployment module includes:

[0039] A model update time monitoring sub-module, which is used to obtain the preset model update time for the target model online service, and monitor whether the current time exceeds the model update time;

[0040] A model update instruction pushing sub-module, which is used to push a model update instruction to the target model online service when it is monitored that the current time exceeds the model update time, and the model update instruction includes the shared storage path.

[0041] In a preferred embodiment, the new model deployment module further includes:

[0042] An update instruction response sub-module, which is used to respond to the model update instruction pushed by the model update service, and obtain the shared storage path included in the instruction;

[0043] A new model acquisition sub-module, configured to obtain the new neural network model from the target service storage space in the target model pool of the shared container service according to the shared storage path;

[0044] A new model online sub-module, configured to, after completing the deployment of the new neural network model, take the current old neural network model offline and bring the new neural network model online to provide an inference function for the current model service.

[0045] To solve the above technical problems, an embodiment of the present application further provides a computer device, including a memory and a processor. Computer-readable instructions are stored in the memory. When the computer-readable instructions are executed by the processor, the processor is caused to execute the steps of the above-mentioned model deployment and update processing method.

[0046] To solve the above technical problems, an embodiment of the present application further provides a storage medium storing computer-readable instructions. When the computer-readable instructions are executed by one or more processors, one or more processors are caused to execute the steps of the above-mentioned model deployment and update processing method.

[0047] To solve the above technical problems, an embodiment of the present application further provides a computer program product, including a computer program and computer instructions. When the computer program and computer instructions are executed by a processor, the processor is caused to execute the steps of the above-mentioned model deployment and update processing method.

[0048] Compared with the prior art, the advantages of the present application are as follows:

[0049] The present application can deploy and manage the neural network models belonging to the online services of each model for the platform automatically. The model deployment management system of the present application includes a model update service and a shared container service. The model update service provides a deployment interface to the development end where the developers are located, so that the developers can upload the newly developed neural network models to the model update service for new model deployment processing, without the need for developers to perform manual deployment, saving the time of developers and improving the model update efficiency.

[0050] Secondly, the shared container service in the present application can manage the neural network models of each version of the model online service in the platform automatically. The shared container service manages the neural network models of each version of each model online service in a pooled and grouped manner to achieve data isolation of each neural network model, ensure the security of model data, and improve the system stability of the model online service.

[0051] In addition, the shared container service in the present application manages the neural network models of different versions of the model online service, so that the platform can deploy multiple versions of models in the shared container service to achieve the operation of multiple versions of models in the model online service. Brief Description of the Drawings

[0052] The above and / or additional aspects and advantages of the present application will become apparent and be readily understood from the following description of embodiments in conjunction with the accompanying drawings, wherein:

[0053] Figure 1 A schematic diagram of a typical network deployment architecture related to implementing the technical solution of the present application;

[0054] Figure 2 A flowchart of a typical embodiment of the method for processing model deployment update of the present application;

[0055] Figure 3 A flowchart formed by the specific implementation manner of the present application regarding storing a new neural network model in the model cloud storage space by the model update service;

[0056] Figure 4 A flowchart formed by the specific implementation manner of the present application regarding the shared container service obtaining a new neural network model from the model cloud storage space for storage processing;

[0057] Figure 5 A flowchart formed by the specific implementation manner of the present application regarding the model update service notifying the target model online service to deploy a new model after the time reaches the model update time;

[0058] Figure 6 A flowchart formed by the specific implementation manner of the present application regarding the target model online service obtaining a new neural network model from the shared container service for hot update;

[0059] Figure 7 A principle block diagram of a typical embodiment of the model deployment update processing device of the present application;

[0060] Figure 8 A basic structure block diagram of a computer device according to an embodiment of the present application. Detailed Description of the Embodiments

[0061] Embodiments of the present application are described in detail below. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described by referring to the accompanying drawings are exemplary only for explaining the present application and should not be construed as limiting the present application.

[0062] Those skilled in the art can understand that, unless specifically stated otherwise, the singular forms "a", "an", "the" and "said" used herein may also include the plural forms. It should be further understood that the term "comprising" used in the specification of the present application means the presence of the stated features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or their groups. It should be understood that when we say that an element is "connected" or "coupled" to another element, it can be directly connected or coupled to other elements, or there may also be intermediate elements. In addition, the "connection" or "coupling" used herein may include wireless connection or wireless coupling. The phrase "and / or" used herein includes all or any unit and all combinations of one or more of the associated listed items.

[0063] Those skilled in the art can understand that, unless otherwise defined, all terms used herein (including technical terms and scientific terms) have the same meaning as the general understanding of those of ordinary skill in the art to which the present application belongs. It should also be understood that terms such as those defined in a general dictionary should be understood to have a meaning consistent with the meaning in the context of the prior art, and will not be interpreted in an idealized or overly formal sense unless specifically defined as herein.

[0064] Those skilled in the art can understand that the "client", "terminal", and "terminal device" used herein include both devices with wireless signal receivers that only have the ability to receive and no ability to transmit, and devices with receiving and transmitting hardware that have the receiving and transmitting hardware capable of two-way communication on a two-way communication link. Such devices may include: cellular or other communication devices such as personal computers, tablet computers, etc., which have a single-line display or a multi-line display or a cellular or other communication device without a multi-line display; PCS (Personal Communications Service), which can combine voice, data processing, fax, and / or data communication capabilities; PDA (Personal Digital Assistant), which may include a radio frequency receiver, a pager, Internet / intranet access, a web browser, a notepad, a calendar, and / or a GPS (Global Positioning System) receiver; traditional laptop and / or palm computers or other devices, which are traditional laptop and / or palm computers or other devices with and / or including a radio frequency receiver. The "client", "terminal", and "terminal device" used herein can be portable, transportable, installed in a vehicle (air, sea, and / or land), or suitable for and / or configured to run locally, and / or run in a distributed manner at any other location on the earth and / or in space. The "client", "terminal", and "terminal device" used herein can also be a communication terminal, an Internet access terminal, a music / video playback terminal, for example, it can be a PDA, a MID (Mobile Internet Device), and / or a mobile phone with music / video playback function, or it can also be a smart TV, a set-top box, etc.

[0065] The hardware referred to by names such as "server", "client", and "working node" in this application is essentially an electronic device with the equivalent capabilities of a personal computer, and is a hardware device with the necessary components disclosed by the von Neumann principle, including a central processing unit (including an arithmetic unit and a controller), a memory, an input device, and an output device. The computer program is stored in its memory, and the central processing unit loads the program stored in the external memory into the memory for execution, executes the instructions in the program, and interacts with the input / output devices to complete specific functions.

[0066] It should be noted that the concept of "server" in this application can similarly be extended to the case applicable to server clusters. According to the network deployment principles understood by those skilled in the art, the servers should be logically divided. Physically, these servers can either be independent of each other but can be invoked through interfaces, or integrated into a physical computer or a set of computer clusters. Those skilled in the art should understand this flexibility and should not be restricted by this when implementing the network deployment method of this application.

[0067] Please refer to Figure 1 , the hardware foundation required for the implementation of the related technical solutions of this application can be deployed according to the architecture shown in the figure. The server 80 referred to in this application is deployed in the cloud. As an online server, it can be responsible for further connecting relevant data servers and other servers providing relevant support, etc., so as to form a logically related service cluster to provide services for relevant terminal devices such as the smartphone 81 and personal computer 82 shown in the figure or a third-party server (not shown). The smartphone and the personal computer can both access the Internet through well-known network access methods and establish a data communication link with the server 80 in the cloud to run the terminal application programs related to the services provided by the server.

[0068] For the server, the application program is usually constructed as a service process, opening corresponding program interfaces for the application programs running on various terminal devices to make remote calls. The related technical solutions suitable for running on the server in this application can be implemented in the server in this way.

[0069] The application program refers to the application program running on the server or the terminal device. This application program implements the related technical solutions of this application in a programming way. Its program code can be saved in a non-volatile storage medium recognizable by the computer in the form of computer-executable instructions and be called into the memory by the central processing unit to run. The related device of this application is constructed through the operation of this application program on the computer.

[0070] For the server, the application program is usually constructed as a service process, opening corresponding program interfaces for the application programs running on various terminal devices to make remote calls. The related technical solutions suitable for running on the server in this application can be implemented in the server in this way.

[0071] Those skilled in the art should be aware that although the various methods of this application are described based on the same concept and thus show commonality with each other, unless otherwise specified, these methods can be executed independently. Similarly, for each of the embodiments disclosed in this application, they are all proposed based on the same inventive concept. Therefore, concepts with the same expression, as well as concepts that are only appropriately transformed for convenience although the concept expressions are different, should be understood equivalently.

[0072] Please refer to Figure 2 , a method for processing model deployment update in this application. In its typical embodiment, it includes the following steps:

[0073] Step S11, in response to a model update event acting on the target model online service, obtain the new neural network model and its corresponding model configuration information included in this event. The model configuration information includes a model pool identifier, a model service identifier, and a model version number:

[0074] The current model update service responds to the model update event acting on the target model online service to obtain the newly updated neural network model included in this model update event and the model configuration information corresponding to this new neural network model.

[0075] The model update event is generally triggered and generated by the model development end. The model development end refers to the development end that establishes a data communication link with the current model update service. After the development user at the model development end completes the new neural network model for the target model online service and edits the model pool identifier, model service identifier, and model version number corresponding to this new neural network model, the model development end will encapsulate the model pool identifier, model service identifier, and model version number to generate the model configuration information, and then trigger the model update event so that the current model update service responds to this model update event to obtain the new neural network model and model configuration information.

[0076] The model update service refers to a service that provides model updates for the model online service, model development end, and shared container service associated with it. Specifically, the model update service is responsible for storing the newly developed or updated neural network model by the model development end in the shared container service, and for the model development end to be responsible for notifying the corresponding model online service to obtain the corresponding new neural network model from the shared container service for deployment; the model online service provides the corresponding neural network model inference service through the neural network model deployed by itself; the shared container service is responsible for storing the newly updated or developed neural network models of each model online service.

[0077] The described model configuration information is used to characterize the model version of the corresponding new neural network model and the corresponding storage space in the shared container service where the new neural network model is stored. For example, the model version included in the model configuration information is used to characterize the model version of the new neural network model, and the model pool identifier and model service identifier included in the model configuration information are used to characterize the storage space of the new neural network model in the shared container service shown. For the specific storage method, please refer to the description of the relevant embodiments in the subsequent steps, and this step will not be elaborated here.

[0078] After the current model update service obtains the new neural network model and the model configuration information, it can store the neural network model and the model configuration information in the model cloud space, so as to drive the shared container service to obtain the neural network model and the model configuration information from the model cloud space and store them in the corresponding storage space, without the current model update service pushing the new neural network model and the model configuration information to the update container service, thus reducing the data transmission pressure of the current model update service.

[0079] Step S12: Push a new model storage instruction to the shared container service, and drive the shared container service to respond to the instruction to store the new neural network model corresponding to the model version number in the target service storage space corresponding to the service identifier in the target model pool corresponding to the model pool identifier according to the model configuration information:

[0080] After the current model update service obtains the new neural network model and the model configuration information, it will push the new model storage instruction to the shared container service, so that the shared container service can respond to the new model storage instruction and store the new neural network model in the corresponding storage space according to the model configuration information.

[0081] The described shared container service is used to share and store the neural network models corresponding to each model online service. There are multiple model pools in the shared container service, and one or more service storage spaces are stored in these model pools. Each service storage space is used to store the neural network models of each version required for the operation of the model online service corresponding to it; the shared container service separates and stores the neural network models required for the operation of each model online service through the multiple model pools and service storage spaces for data isolation, avoiding data overwriting each other and improving the system stability of each model online service.

[0082] After the shared container service responds to the new model storage instruction pushed by the current model update service, it will obtain the new neural network model and its corresponding configuration information through the data communication link established with the current model update service, or obtain the new neural network model and its corresponding configuration information pushed by the current model update service. Additionally, when the current model update service stores the new neural network model and its corresponding configuration information in the model cloud space, after the shared container service responds to the new model storage instruction, it will correspondingly obtain the new neural network model and its corresponding configuration information from this model cloud space.

[0083] Regarding the specific manner in which the shared container service stores and processes the new neural network model according to the model configuration information, after the shared container service obtains the new neural network model and its corresponding model configuration information, it will query the target model pool corresponding to the model pool identifier in multiple model pools according to the model pool identifier included in the model configuration information, and query the target service storage space corresponding to the service identifier in multiple service storage spaces in the target model pool according to the service identifier included in the model configuration information, so as to store the new neural network model and the model version number as mapping relationship data in the target service storage space.

[0084] Step S13, when the target model online service meets the preset new model deployment policy, push the shared storage path corresponding to the new neural network model to the target model online service, and drive the target model online service to obtain the new neural network model from the shared container service according to the shared storage path for deployment:

[0085] After the shared container service completes the storage and processing of the new neural network model, the current model update service will monitor the target model online service. When it monitors that the target model online service meets the preset new model deployment processing, it will drive the target model online service to obtain the new neural network model from the shared container service for deployment.

[0086] The new model deployment strategy described above is a strategy used by the current model update service to determine whether to drive the target model online service to deploy a new neural network model. The current model update service can determine whether the current time exceeds the model update time preset for the target model online service. If it exceeds, it drives the target model online service to deploy a new neural network model for going online. Or the current model update service determines whether the workload of the target model online service requires a new neural network model to be deployed online to improve its service efficiency. Or the current update service determines whether the storage capacity of the target model online service can meet the storage requirements of the new neural network model. If it meets, the target model online service deploys a new neural network model for going online. Regarding the new model deployment strategy, those skilled in the art can flexibly design it according to the business scenario of the model online service, and then deploy the developed new model deployment strategy into the current model update service, and then the current model update service makes a strategy judgment.

[0087] When the current model update service determines that the target model online service meets the preset new model deployment strategy, it will push the shared storage path pointing to the new neural network model pushed by the shared container service to the target model online service, so that the target model pointing service can obtain the new neural network model and its corresponding model version number from the shared container service according to this shared storage path. The shared container service stores the new neural network model and its corresponding model version number in the target service storage space correspondingly, generates a shared storage path representing the new neural network model from the target model pool to the target service storage space, and pushes this shared storage path to the current model update service.

[0088] After obtaining the shared storage path, the target model online service will, according to the shared storage path, obtain the new neural network model and its corresponding model version number from the target service storage space of the target model pool pointed to by this shared storage path in the shared container service, and then deploy this new neural network model to take the place of the old neural network model currently online, and then go online the new neural network model to provide model inference services based on this new neural network model for the client or server associated with this target model online service.

[0089] The current model sharing service can push the shared storage path to multiple model online services applicable to the new neural network model, so that these model online services can obtain the new neural network model from the shared container service for deployment.

[0090] As can be seen from the typical implementation of this method, this method can automate the deployment and management of the neural network models belonging to the online services of each model for the platform. The model deployment and management system of this application includes a model update service and a shared container service. The model update service provides a deployment interface to the development end where the developers are located, so that the developers can upload the newly developed neural network models to the model update service for new model deployment processing, without the need for developers to perform manual deployment, saving the time of developers and improving the model update efficiency. Secondly, the shared container service in this method can automate the management of the neural network models of each version of the model online service for the platform. The shared container service manages the neural network models of each version of each model online service in a pooled and grouped manner to achieve data isolation of each neural network model, ensure the security of model data, and improve the system stability of the model online service. In addition, the shared container service in this method manages the neural network models of different versions of the model online service, so that the platform can deploy multiple versions of the model in the shared container service to achieve the operation of multiple versions of the model for the model online service.

[0091] The above typical embodiments and their variant embodiments fully disclose the implementation scheme of the model deployment and update processing method of this application. However, various variant embodiments of this method can still be deduced by transforming and amplifying some technical means. Other embodiments are outlined as follows:

[0092] In one embodiment, please refer to Figure 3 , in the step of obtaining the new neural network model and its corresponding model configuration information included in the event in response to the model update event of the target model online service, it includes:

[0093] Step S111, receiving the new neural network model and its corresponding model configuration information pushed by the model development end through the model upload interface:

[0094] The current model update service receives the new neural network model and the corresponding model configuration information pushed by the model development end through the model upload interface. The model upload interface is a data communication link interface provided by the current model update service for the model development end.

[0095] Step S112, storing the new neural network model and the model configuration information correspondingly in the model cloud storage space:

[0096] After the current model update service obtains the new neural network model and the model configuration information, it stores the new neural network model and the model configuration information in the cloud storage space, so that the shared container service can obtain the new neural network model and the model configuration information from the cloud storage space for storage processing of the new neural network model. Multiple neural network models and corresponding model configuration information are stored in the cloud storage space, so that the associated shared container service can obtain the corresponding new neural network model for storage processing.

[0097] In this embodiment, the current model update service stores the newly developed neural network model and the corresponding model configuration information in the cloud space to handle the model forwarding of the update container service, effectively reducing the data transmission pressure of the current model update service and saving local storage space.

[0098] In one embodiment, please refer to Figure 4 , in the step of driving the shared container service to respond to the instruction to store the new neural network model corresponding to the model version number in the target service storage space corresponding to the service identifier in the target model pool corresponding to the model pool identifier according to the model configuration information, the following steps executed by the shared container service are included:

[0099] Step S121, in response to the new model storage instruction pushed by the model update service, obtain the new neural network model and its model configuration information corresponding to the instruction from the model cloud storage space:

[0100] The shared container service responds to the new model storage instruction pushed by the model update service, determines the new neural network model pointed to by the new storage instruction, and then obtains the new neural network model and its corresponding model configuration information through the data communication link established with the model cloud storage space.

[0101] Step S122, obtain the model pool identifier and the model service identifier included in the model configuration information, and query the target service storage space corresponding to the service identifier in the target model pool corresponding to the model pool identifier:

[0102] The shared container service parses the model configuration information, obtains the model pool identifier and the model service identifier included in the model configuration information, determines the target model pool corresponding to the model pool identifier, and then queries the target service storage space corresponding to the model service identifier in the target model pool.

[0103] Step S123, obtain the model version number included in the model configuration information, and store the new neural network model and the model version number correspondingly in the target service storage space:

[0104] After determining the target service storage space, the shared container service stores the new neural network model and the model version number included in the model configuration information in the target service storage space correspondingly.

[0105] In this embodiment, the shared container service obtains the new neural network model and the model configuration information from the cloud storage space for storing the new neural network model, isolates the model of the new neural network model, prevents data pollution and data overwriting of the new neural network model, ensures the security of model data, and is convenient for obtaining a new model for the corresponding model online service.

[0106] In one embodiment, please refer to Figure 5 , in the step of pushing the shared storage path corresponding to the new neural network model to the target model online service when the target model online service meets the preset new model deployment strategy, it includes:

[0107] Step S131, obtain the model update time preset for the target model online service, and monitor whether the current time exceeds the model update time:

[0108] The current model update service monitors the target model online service, obtains the model update time preset for the target model online service in advance, and determines whether the current time exceeds the model update time.

[0109] Step S132, when it is monitored that the current time exceeds the model update time, push a model update instruction to the target model online service, and the model update instruction includes the shared storage path:

[0110] When the current model update service monitors that the current time exceeds the model update time, it will push a model update instruction including the shared storage path to the target model online service, so that the target online service responds to the instruction, and then obtains the new neural network model and its model version number from the shared container service according to the shared storage path for model deployment.

[0111] In this embodiment, the model sharing service only drives the corresponding model online service to deploy a new neural network model after the time reaches the model update time, so as to effectively deploy and update the new model and ensure the current operation stability of the model online service.

[0112] In one embodiment, please refer to Figure 6 , in the step of driving the target model online service to obtain a new neural network model from the shared container service according to the shared storage path for deployment, it includes the following steps executed by the target model online service:

[0113] Step S131’, in response to the model update instruction pushed by the model update service, obtain the shared storage path included in the instruction:

[0114] After the target model online service receives the model update instruction pushed by the model update service, it will obtain the shared storage path included in the instruction.

[0115] Step S132’, according to the shared storage path, obtain the new neural network model from the target service storage space in the target model pool of the shared container service:

[0116] After the target model online service obtains the shared storage path, it will obtain the new neural network model and its corresponding model version number from the target service storage space of the target model pool pointed to by the shared storage path in the shared container service according to the shared storage path.

[0117] Step S133’, after completing the deployment of the new neural network model, take the current old neural network model offline and bring the new neural network model online to provide an inference function for the current model service:

[0118] The target model online service obtains the new neural network model and its corresponding model version number, and will deploy the new neural network model to take the currently online old neural network model offline, and then bring the new neural network model online to provide a model inference service based on the new neural network model for the client or server associated with the target model online service.

[0119] In this embodiment, the model online service can obtain a new model from the shared container for deployment, and then perform model hot update to provide a new version of the model inference service, without having to obtain the new model from the model update service until then, so as to save the data transmission pressure of the model update service and improve the overall execution efficiency of the model hot update system.

[0120] Furthermore, by functionalizing each step in the methods disclosed in the above embodiments, a model deployment and update processing device of the present application can be constructed. In accordance with this idea, please refer to Figure 7, in one typical embodiment, the device includes: an update event response module 11, configured to respond to a model update event acting on the target model online service, and obtain the new neural network model and its corresponding model configuration information included in the event, where the model configuration information includes a model pool identifier, a model service identifier, and a model version number; a model shared storage module 12, configured to push a new model storage instruction to the shared container service, and drive the shared container service to respond to the instruction to store the new neural network model corresponding to the model version number in the target service storage space corresponding to the service identifier in the target model pool corresponding to the model pool identifier according to the model configuration information; a new model deployment module 13, configured to push the shared storage path corresponding to the new neural network model to the target model online service when the target model online service meets the preset new model deployment policy, and drive the target model online service to obtain the new neural network model from the shared container service according to the shared storage path for deployment.

[0121] In one embodiment, the update event response module 11 includes: a model acquisition sub-module, configured to receive the new neural network model and its corresponding model configuration information pushed by the model development end through the model upload interface; a model cloud storage sub-module, configured to store the new neural network model and the model configuration information correspondingly in the model cloud storage space.

[0122] In one embodiment, the model shared storage module 12 includes: a storage instruction response sub-module, configured to respond to the new model storage instruction pushed by the model update service, and obtain the new neural network model and its model configuration information corresponding to the instruction from the model cloud storage space; a storage space query sub-module, configured to obtain the model pool identifier and the model service identifier included in the model configuration information, and query the target service storage space corresponding to the service identifier in the target model pool corresponding to the model pool identifier; a model storage sub-module, configured to obtain the model version number included in the model configuration information, and store the new neural network model and the model version number correspondingly in the target service storage space.

[0123] In one embodiment, the new model deployment module 13 includes: a model update time monitoring sub-module, configured to obtain the preset model update time for the target model online service, and monitor whether the current time exceeds the model update time; a model update instruction push sub-module, configured to push a model update instruction to the target model online service when it is monitored that the current time exceeds the model update time, where the model update instruction includes the shared storage path.

[0124] In another embodiment, the new model deployment module 13 further includes: an update instruction response sub-module, configured to respond to a model update instruction pushed by a model update service and obtain a shared storage path included in the instruction; a new model acquisition sub-module, configured to obtain the new neural network model from the target service storage space in the target model pool of the shared container service according to the shared storage path; and a new model online sub-module, configured to, after completing the deployment of the new neural network model, take the current old neural network model offline and bring the new neural network model online to provide an inference function for the current model service.

[0125] To solve the above technical problems, an embodiment of the present application further provides a computer device for running a computer program implemented according to the model deployment update processing method. For details, please refer to Figure 8 , Figure 8 which is a basic structural block diagram of the computer device in this embodiment.

[0126] As Figure 8 shown, it is a schematic internal structure diagram of the computer device. The computer device includes a processor, a non-volatile storage medium, a memory, and a network interface connected through a system bus. Among them, the non-volatile storage medium of the computer device stores an operating system, a database, and computer-readable instructions. Control information sequences can be stored in the database. When the computer-readable instructions are executed by the processor, the processor can implement a model deployment update processing method. The processor of the computer device is used to provide computing and control capabilities to support the operation of the entire computer device. Computer-readable instructions can be stored in the memory of the computer device. When the computer-readable instructions are executed by the processor, the processor can execute a model deployment update processing method. The network interface of the computer device is used to connect and communicate with a terminal. Those skilled in the art can understand that Figure 8 the structure shown in

[0127] is only a block diagram of some structures related to the solution of the present application and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0128] The present application also provides a non-volatile storage medium. The model deployment update processing method is written as a computer program and stored in the storage medium in the form of computer-readable instructions. When the computer-readable instructions are executed by one or more processors, it means the program runs on the computer, thereby enabling one or more processors to execute the steps of the model deployment update processing method in any of the above embodiments.

[0129] Those of ordinary skill in the art can understand that all or part of the processes of implementing the methods in the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a magnetic disk, an optical disk, a read-only memory (ROM), etc., or a random access memory (RAM), etc.

[0130] In summary, the present application realizes the automated management of the deployment and update of a new neural network model, saves the deployment time of the new model, improves the model hot update efficiency, and isolates and manages the model data to ensure its data security.

[0131] It should be understood that although the steps in the flowchart of the accompanying drawings are shown in sequence according to the indication of the arrows, these steps do not necessarily have to be executed in the order indicated by the arrows. Unless there is a clear indication in this article, the execution of these steps has no strict order limit and can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages. These sub-steps or stages do not necessarily have to be executed at the same time, but can be executed at different times. Their execution order does not necessarily have to be sequential, but can be executed alternately or alternately with at least a part of other steps or sub-steps or stages of other steps.

[0132] Those skilled in the art of this technology can understand that the various operations, methods, steps, measures, and solutions in the processes discussed in the present application can be alternated, changed, combined, or deleted. Further, other steps, measures, and solutions in the various operations, methods, and processes discussed in the present application can also be alternated, changed, rearranged, decomposed, combined, or deleted. Further, the steps, measures, and solutions in the prior art that are the same as those disclosed in the present application can also be alternated, changed, rearranged, decomposed, combined, or deleted.

[0133] The above are only some embodiments of the present application. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present application, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present application.

Claims

1. A method for processing model deployment updates, characterized in that, Including the following steps: Responding to a model update event acting on the target model online service, obtaining the new neural network model and its corresponding model configuration information included in the event, where the model configuration information includes a model pool identifier, a model service identifier, and a model version number; Pushing a new model storage instruction to the shared container service, driving the shared container service to respond to the instruction to store the new neural network model and the model version number correspondingly in the target service storage space corresponding to the model service identifier in the target model pool corresponding to the model pool identifier according to the model configuration information; When the target model online service meets the preset new model deployment policy, pushing the shared storage path corresponding to the new neural network model to the target model online service, driving the target model online service to obtain the new neural network model from the shared container service according to the shared storage path for deployment, including the following steps executed by the target model online service: Responding to a model update instruction pushed by the model update service, obtaining the shared storage path included in the instruction; According to the shared storage path, obtaining the new neural network model from the target service storage space in the target model pool corresponding to the shared container service; After completing the deployment of the new neural network model, taking offline the current old neural network model and bringing online the new neural network model to provide an inference function for the current model service.

2. The method according to claim 1, characterized in that, In the step of responding to a model update event acting on the target model online service and obtaining the new neural network model and its corresponding model configuration information included in the event, it includes: Receiving the new neural network model and its corresponding model configuration information pushed by the model development side through the model upload interface; Correspondingly storing the new neural network model and the model configuration information in the model cloud storage space.

3. The method according to claim 2, wherein In the step of driving the shared container service to respond to the instruction to store the new neural network model and the model version number correspondingly in the target service storage space corresponding to the model service identifier in the target model pool corresponding to the model pool identifier according to the model configuration information, it includes the following steps executed by the shared container service: Responding to a new model storage instruction pushed by the model update service, obtaining the new neural network model and its model configuration information corresponding to the instruction from the model cloud storage space; Obtaining the model pool identifier and the model service identifier included in the model configuration information, and querying the target service storage space corresponding to the service identifier in the target model pool corresponding to the model pool identifier; Obtaining the model version number included in the model configuration information, and correspondingly storing the new neural network model and the model version number in the target service storage space.

4. The method according to claim 1, wherein In the step of pushing the shared storage path corresponding to the new neural network model to the target model online service when the target model online service meets the preset new model deployment policy, it includes: Obtaining the preset model update time for the target model online service, and monitoring whether the current time exceeds the model update time; When it is monitored that the current time exceeds the model update time, a model update instruction is pushed to the target model online service, and the model update instruction includes the shared storage path.

5. The method according to claim 1, wherein In the step of pushing the shared storage path corresponding to the new neural network model to the target model online service, the shared storage path is pushed by the shared container service to the current model update service, and the shared storage path points to the new neural network model in the target service storage space of the target model pool.

6. A model deployment update processing device, characterized in that It includes: An update event response module, configured to respond to a model update event acting on the target model online service, and obtain the new neural network model and its corresponding model configuration information included in the event, where the model configuration information includes a model pool identifier, a model service identifier, and a model version number; A model shared storage module, configured to push a new model storage instruction to the shared container service, and drive the shared container service to respond to the instruction to store the new neural network model and the model version number in the target service storage space corresponding to the model service identifier in the target model pool corresponding to the model pool identifier according to the model configuration information; A new model deployment module, configured to push the shared storage path corresponding to the new neural network model to the target model online service when the target model online service meets a preset new model deployment strategy, and drive the target model online service to obtain the new neural network model from the shared container service according to the shared storage path for deployment.

7. The device according to claim 6, characterized in that, The new model deployment module further includes: An update instruction response sub-module, configured to respond to a model update instruction pushed by the model update service, and obtain the shared storage path included in the instruction; A new model acquisition sub-module, configured to obtain the new neural network model from the target service storage space in the target model pool corresponding to the shared container service according to the shared storage path; A new model online sub-module, configured to, after completing the deployment of the new neural network model, take offline the current old neural network model and bring online the new neural network model to provide an inference function for the current model service.

8. An electronic device, comprising a central processing unit and a memory, characterized in that, The central processing unit is used to call and run the computer program stored in the memory to execute the steps of the method according to any one of claims 1 to 5.

9. A non-volatile storage medium, characterized in that, It stores a computer program implemented according to the method according to any one of claims 1 to 5 in the form of computer-readable instructions, and when the computer program is called and run by the computer, it executes the steps included in the method.

10. A computer program product, comprising a computer program / instructions, characterized in that, When the computer program / instructions are executed by the processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Images