Data Processing Methods and Systems for an Artificial Intelligence Experimental Cloud Platform Based on Kubernetes

Abstract

This application relates to cloud platform technology and provides a data processing method and system for an AI experimental cloud platform based on Kubernetes. The method involves the AI ​​experimental cloud platform verifying access requests from target user terminals and then obtaining the corresponding target container instance. The target container instance then receives target data sent by the target user terminal, performs corresponding model calculations to obtain target operation data, and stores it. Finally, the target user terminal with editing permissions retrieves the target operation data, edits it, and obtains the corresponding target edited data, which is then sent back to the AI ​​experimental cloud platform. This approach enables the construction of AI-related basic frameworks and algorithm models within the containers of the AI ​​experimental cloud platform, allowing for the direct processing of AI experimental tasks within the containers without requiring improvements to the local hardware performance of the user terminal. Furthermore, it avoids manually building AI-related basic frameworks and algorithm models on the terminal, thus improving the processing efficiency of AI experimental tasks.

Description

Technical Field

[0001] This application relates to the field of cloud platform technology, and in particular to a data processing method and system for an artificial intelligence experimental cloud platform based on K8S. Background Technology

[0002] Currently, when enterprises or universities conduct AI-related experiments, participants first need to manually build the basic AI framework and complex algorithm models on their devices before they can adjust the algorithm parameters. This process is problematic because: firstly, the devices used by participants generally lack the performance requirements for large-scale AI experiments; secondly, manually building the basic AI framework and complex algorithm models on these devices requires a high level of expertise and is time-consuming and labor-intensive. This results in inadequate hardware performance and low computational efficiency in building the basic AI framework for AI experiments. Summary of the Invention

[0003] This application provides a data processing method and system for an artificial intelligence experimental cloud platform based on K8S, aiming to solve the problems of insufficient hardware performance and low computational efficiency in building artificial intelligence-related basic frameworks in existing technologies.

[0004] In a first aspect, embodiments of this application provide a data processing method for an artificial intelligence experimental cloud platform based on Kubernetes (K8S), applied to an artificial intelligence experimental cloud platform system. The artificial intelligence cloud platform system includes at least one user terminal and an artificial intelligence experimental cloud platform itself. The artificial intelligence experimental cloud platform includes a Web application layer, a cloud platform interface layer, a cluster management layer, a resource container layer, and a physical hardware layer. The method includes:

[0005] The target user terminal sends a container access request to the artificial intelligence experimental cloud platform; wherein, the target user terminal is any one of the at least one user terminal;

[0006] The artificial intelligence experimental cloud platform verifies the access request. If the access request passes the verification, it obtains the target container instance corresponding to the access request.

[0007] The target user terminal is connected to the target container instance in the artificial intelligence experimental cloud platform;

[0008] The target container instance in the artificial intelligence experimental cloud platform receives the target data sent by the target user terminal, and stores the target operation data obtained by model calculation and the target data in the storage space corresponding to the physical hardware layer in the artificial intelligence experimental cloud platform.

[0009] The target grading permission user terminal obtains the target operation data, obtains the corresponding target grading data based on the target operation data, and sends the target grading data to the artificial intelligence experimental cloud platform; wherein, the target grading permission user terminal is a user terminal with grading permission in the target group to which the target user terminal belongs among the at least one user terminal.

[0010] Secondly, this application provides a data processing system for an artificial intelligence experimental cloud platform based on K8S, which includes at least one user terminal and an artificial intelligence experimental cloud platform; the artificial intelligence experimental cloud platform includes a web application layer, a cloud platform interface layer, a cluster management layer, a resource container layer, and a physical hardware layer;

[0011] The target user terminal is used to send the container access request to the artificial intelligence experimental cloud platform; wherein the target user terminal is any one of the at least one user terminal.

[0012] The artificial intelligence experimental cloud platform is used to verify the access request. If the access request passes the verification, the target container instance corresponding to the access request is obtained.

[0013] The target user terminal is also used to connect to the target container instance in the artificial intelligence experimental cloud platform;

[0014] The artificial intelligence experimental cloud platform is also used to receive target data sent by the target user terminal through the target container instance, and store the target operation data obtained by model calculation of the target data and the target data in the storage space corresponding to the physical hardware layer in the artificial intelligence experimental cloud platform;

[0015] A target grading permission user terminal is used to acquire the target operation data, obtain corresponding target grading data based on the target operation data, and send the target grading data to the artificial intelligence experimental cloud platform; wherein, the target grading permission user terminal is a user terminal with grading permission in the target group to which the target user terminal belongs among the at least one user terminal.

[0016] Thirdly, embodiments of this application also provide a computer device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the data processing method of the K8S-based artificial intelligence experimental cloud platform described in the first aspect.

[0017] Fourthly, embodiments of this application also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the data processing method of the K8S-based artificial intelligence experimental cloud platform described in the first aspect.

[0018] This application provides a data processing method and system for an AI experimental cloud platform based on Kubernetes. After the AI ​​experimental cloud platform verifies the access request from a target user terminal, it obtains the corresponding target container instance. The target container instance then receives and stores the target operation data sent by the target user terminal. Finally, the user terminal with the target editing permission obtains the target operation data, edits it, and sends the corresponding edited data back to the AI ​​experimental cloud platform. This allows the AI-related basic framework and algorithm models to be pre-built within the container of the AI ​​experimental cloud platform. Users can directly complete AI experimental tasks within the container without needing to upgrade the local hardware performance of the user terminal. Furthermore, it avoids manually building the AI-related basic framework and algorithm models on the user terminal, thus improving the processing efficiency of AI experimental tasks. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 A schematic diagram illustrating an application scenario of the data processing method for an artificial intelligence experimental cloud platform based on K8S provided in this application embodiment;

[0021] Figure 2 A flowchart illustrating the data processing method for an AI experimental cloud platform based on K8S provided in this application embodiment;

[0022] Figure 3 A schematic block diagram of a data processing system for an artificial intelligence experimental cloud platform based on K8S, provided in an embodiment of this application;

[0023] Figure 4 A schematic block diagram of a computer device provided in an embodiment of this application. Detailed Implementation

[0024] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0025] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0026] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the scope of the application. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0027] It should also be further understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0028] Please see Figure 1 and Figure 2 , Figure 1 A schematic diagram illustrating an application scenario of the data processing method for an artificial intelligence experimental cloud platform based on K8S provided in this application embodiment; Figure 2 This is a flowchart illustrating the data processing method for an AI experimental cloud platform based on Kubernetes (K8S) provided in this application embodiment. The K8S-based data processing method for an AI experimental cloud platform provided in this application embodiment is applied to an AI experimental cloud platform system, such as... Figure 1 As shown, the artificial intelligence cloud platform system includes at least one user terminal and an artificial intelligence experimental cloud platform; and the artificial intelligence experimental cloud platform includes a web application layer, a cloud platform interface layer, a cluster management layer, a resource container layer, and a physical hardware layer.

[0029] The Web application layer not only provides users with different ways to log in to the AI ​​experiment cloud platform, but also offers an online visual interface. When a user logs in to the AI ​​experiment cloud platform via their user terminal, the first thing displayed is the Web application layer. Through this layer, users can perform operations such as resource application, task management, model management, resource viewing, experiment management, experiment viewing, experiment submission, and logging into the container interaction interface. The visual interface (i.e., the personal container interaction interface) within the Web application layer primarily uses the Jupyter-Notebook service, allowing users to directly write and run AI experiment-related code on their user terminal using the corresponding web client.

[0030] The cloud platform interface layer is used for deploying and developing specific related business applications. For example, the cluster management layer uses native APIs (API stands for Application Program Interface), and the cloud platform interface layer encapsulates these native APIs to provide request access interfaces to the upper-layer web application layer. The cloud platform interface layer can also receive user requests from user terminals to provide corresponding services to the users. Functionally, the cloud platform interface layer is mainly divided into: a user management module (for users with administrator privileges to upload grade information, and manage user login, registration, information modification, etc.); a resource management module (for applying for container instances using GPU resources, or building images to a private image repository, accessing personal container instances, etc.); a cluster management module (for obtaining the running status of container instances, server running status, etc.); and an experiment management module (for setting up experiments, grading, distributing teaching materials, etc.).

[0031] The cluster management layer is the core of the entire AI experimental cloud platform, implemented based on components such as Kubernetes (K8S) and Docker. Specifically, it achieves unified orchestration and scheduling management of containers through the Kubernetes API. The cluster management layer can receive user terminal operations to select image creation dependency descriptions and construct script files, and then execute them to generate new container instances. Kubernetes is a container tool that provides guarantees for the normal operation of containers and manages the nodes on those containers; it is used for unified management and control of the lifecycle and resource allocation of containers and application layer services; it is also used for real-time monitoring and management of containers and application services deployed within them; it is responsible for unified Pod scheduling for tasks such as cluster monitoring, image creation and management, and image repository management on the AI ​​experimental cloud platform; it is also responsible for the scheduling and management of computing resources, and can automatically create cloud clusters according to the needs of different AI experimental tasks and the different experimental parameters set by teachers.

[0032] The resource container layer supports the runtime environment of the entire AI experimental cloud platform. It is implemented based on components such as Docker and a private image repository, and provides on-demand allocation of resources such as CPU, GPU, and storage. The Docker container engine within the resource container layer provides a container runtime environment and can create images for different needs. The private image repository integrates framework images such as TensorFlow, Caffe, and PyTorch required for AI-related experiments, and also supports deep neural networks (DNN), convolutional neural networks (CNN), and object detection-related YoLoV1–V5 models.

[0033] The physical hardware layer, located at the bottom layer of the entire AI experimental cloud platform, provides the various underlying infrastructure and hardware devices (such as CPUs, GPUs, hard drives, server nodes, etc.) required by the AI ​​experimental cloud platform. These underlying infrastructure and hardware devices are uniformly managed and scheduled by Kubernetes, allocated as container runtime resources according to actual needs, and provide hardware support for model training and inference.

[0034] like Figure 2 As shown, the data processing method of the K8S-based artificial intelligence experimental cloud platform includes steps S101 to S105.

[0035] S101, The target user terminal sends a container access request to the artificial intelligence experimental cloud platform; wherein, the target user terminal is any one of the at least one user terminal.

[0036] In this embodiment, when a user logs into the AI ​​experiment cloud platform (which can be considered a container cluster using Kubernetes container orchestration technology or a cloud server cluster) using a user terminal such as a cloud terminal, smartphone, PC (mainly desktop computer), or tablet, the user sends a container access request to the AI ​​experiment cloud platform using the target user terminal. Specifically, the user can connect to the AI ​​experiment cloud platform based on the integrated development environment (such as VS Code or PyCharm) on the target user terminal, thereby further entering the container corresponding to that user to perform AI experiment tasks. Since there is no need to build an AI experiment environment locally on the user terminal, but instead directly logs into the corresponding container in the AI ​​experiment cloud platform to conduct AI-related experiments, the time required to build an AI experiment environment locally is reduced, and it also eliminates the need for users with high levels of professional knowledge to build a local AI experiment environment.

[0037] In one embodiment, step S101 includes:

[0038] The virtual terminal of the Web application layer displayed in the target user terminal establishes a communication connection with the target container instance in the artificial intelligence experimental cloud platform through the WebSocket protocol.

[0039] In this embodiment, the Web application layer in the AI ​​experimental cloud platform can be regarded as a virtual terminal simulated by a JavaScript library. This virtual terminal is displayed on the target user terminal. The virtual terminal of the Web application layer displayed on the target user terminal establishes a communication connection with the target container instance in the AI ​​experimental cloud platform through the WebSocket protocol to maintain state and perform front-end and back-end interaction. Therefore, this is an implementation method of dynamically connecting a personal container instance via an AI experimental cloud platform.

[0040] Of course, if a user connects to the AI ​​experimental cloud platform using an integrated development environment (such as VS Code or PyCharm) on the target user's terminal, it needs to determine whether the user has selected a base image with Jupyter components (specifically, Jupyter-Notebook components). If the user has not selected a base image with Jupyter components, the selected base image will be automatically built into an image that supports SSH protocol access (a secure protocol built on the application and transport layers). Specifically, the AI ​​experimental cloud platform backend will automatically configure a Dockerfile script file, using the user's login account and password as the login account and password for the remote SSH service, and mapping the container's port to the host machine's port. Furthermore, if the user uses VS Code to connect to the AI ​​experimental cloud platform, the user can access the corresponding container instance in the AI ​​experimental cloud platform through the open-ssh plugin in VS Code by accessing the host machine's IP address and port.

[0041] In one embodiment, the method further includes the following steps before step S101:

[0042] If the artificial intelligence experimental cloud platform receives a user account creation instruction from a user with administrator privileges, it obtains a list of user accounts corresponding to the user account creation instruction.

[0043] The cluster management layer in the artificial intelligence experimental cloud platform creates namespaces corresponding to the user account list.

[0044] In this embodiment, at least three types of user accounts with different permissions are pre-defined in the AI ​​experimental cloud platform: administrator user accounts, first-permission user accounts (such as teacher user accounts), and second-permission user accounts (such as student user accounts). Administrator user accounts have the authority to manage all data within the entire AI experimental cloud platform; for example, creating multiple namespaces in the cluster management layer of the AI ​​experimental cloud platform is one such permission. First-permission user accounts have the authority to create multiple containers within their corresponding namespaces for second-permission user accounts to log in and use. Second-permission user accounts only have the authority to log in to the corresponding containers within their respective namespaces to process AI experimental tasks.

[0045] In this system, an administrator-privileged user account can receive a list of user accounts to be created from a user corresponding to a first-privilege user account. The administrator, after logging into the AI ​​Experiment Cloud Platform, can then combine the teacher's name and the student's class name from the user's bill list to create a combined name. The administrator then creates a namespace within the cluster management layer of the AI ​​Experiment Cloud Platform using this combined name. This can be visualized as the administrator creating a class-specific namespace for the class taught by the teacher whose name corresponds to that teacher's name on the AI ​​Experiment Cloud Platform. Furthermore, the administrator can also match the number of second-privilege user accounts (i.e., the total number of student names in the user's bill list is the same as the total number of second-privilege user accounts) to the student name list (or student ID list) included in the user's bill list.

[0046] Of course, when creating multiple containers within each namespace according to actual needs, it could be that an administrator-privileged user account creates a corresponding number of containers based on the requirements of the AI ​​experiment task and the student name list included in the user billing list; or it could be that a teacher with first-privilege user accounts creates a corresponding number of containers based on the requirements of the AI ​​experiment task and the student name list included in the user billing list. In the AI ​​experiment cloud platform, the relevant information for the namespaces is stored in the cluster management layer, and the containers corresponding to each namespace are deployed in the resource container layer of the AI ​​experiment cloud platform.

[0047] In one embodiment, after the cluster management layer in the artificial intelligence experimental cloud platform creates a namespace corresponding to the user account list, it further includes:

[0048] If the resource container layer in the artificial intelligence experimental cloud platform detects a unified experimental task deployment instruction, it obtains the first target image resource, GPU resource, and data storage volume path corresponding to the unified experimental task deployment instruction. The resource container layer then creates a container based on the first target image resource, GPU resource, and data storage volume path corresponding to the unified experimental task deployment instruction.

[0049] If the resource container layer in the artificial intelligence experimental cloud platform detects a personalized container deployment instruction, it obtains a second target image resource corresponding to the personalized container deployment instruction. The resource container layer then creates a container based on the second target image resource corresponding to the personalized container deployment instruction and the pre-stored data storage volume path.

[0050] In this embodiment, after the namespace for a specific class, such as Class A, is created in the cluster management layer of the artificial intelligence experimental cloud platform, the teacher corresponding to the first-authority user account can generate a unified experimental task deployment instruction based on the requirements of the artificial intelligence experimental task. This unified experimental task deployment instruction specifically sets the first target image resource, GPU resource, and data storage volume path. Subsequently, the resource container layer creates containers corresponding to the first target image resource, GPU resource, and data storage volume path specified in the unified experimental task deployment instruction.

[0051] The first target image resource can be selected from one or more of the following: images integrating frameworks such as TensorFlow, Caffe, and PyTorch required for experiments in the field of artificial intelligence; or images integrating artificial intelligence-related neural network models such as deep neural networks (DNN), convolutional neural networks (CNN), and object detection networks such as YoLoV1 to V5. More specifically, the first target image resource can be selected from the TensorFlow framework, and a convolutional neural network (CNN) can be deployed on the TensorFlow framework.

[0052] Once the teacher corresponding to the first-authority user account generates a unified experimental task deployment instruction based on the requirements of the artificial intelligence experimental task, and completes the creation of multiple containers corresponding to the namespace in the resource container layer, the initial environment setup for the artificial intelligence experimental task is completed.

[0053] Of course, after the namespace for a class, such as Class A, is created in the cluster management layer of the AI ​​experiment cloud platform, the student corresponding to the second-authority user account can generate a personalized container deployment instruction based on their individual needs for AI experiment tasks. This personalized container deployment instruction is then sent by the cloud platform interface layer of the AI ​​experiment cloud platform to the user terminal used by the teacher corresponding to the first-authority user account. Once the teacher approves the personalized container deployment instruction on the user terminal, the resource container layer of the AI ​​experiment cloud platform creates a container based on the second target image resource corresponding to the personalized container deployment instruction and the pre-stored data storage volume path. Similarly, the second target image resource can be selected from one or more of the following: images integrating frameworks such as TensorFlow, Caffe, and PyTorch required for AI-related experiments; or images integrating AI-related neural network models such as Deep Neural Networks (DNN), Convolutional Neural Networks (CNN), and object detection networks such as YoLoV1 to V5. More specifically, the second target image resource can select the TensorFlow framework and deploy the YoLoV5 object detection network on the TensorFlow framework.

[0054] Furthermore, when a student with the second-level user account generates a personalized container deployment instruction based on their individual needs for conducting AI experiments, GPU resources are not allocated by default. This means that containers created based on the personalized deployment instruction are ordinary server containers, not GPU server containers. However, if the personalized deployment instruction requires the use of a GPU server container, the instruction is sent to the user terminal used by the teacher with the first-level user account. Once the teacher approves the personalized deployment instruction on their user terminal, the resource container layer of the AI ​​experiment cloud platform creates a container based on the second target image resources, GPU resources, and pre-stored data storage volume paths corresponding to the personalized deployment instruction.

[0055] In one embodiment, if the resource container layer in the artificial intelligence experimental cloud platform detects a unified experimental task deployment instruction, it obtains the first target image resource, GPU resource, and data storage volume path corresponding to the unified experimental task deployment instruction. The resource container layer then creates a container based on the first target image resource, GPU resource, and data storage volume path corresponding to the unified experimental task deployment instruction, including:

[0056] If the resource container layer in the artificial intelligence experimental cloud platform detects a primary experimental task deployment instruction corresponding to a unified experimental task deployment instruction, it obtains the first sub-target image resource, GPU resource, and data storage volume path corresponding to the primary experimental task deployment instruction. The resource container layer then creates a container based on the first sub-target image resource, first GPU resource, and first data storage volume path corresponding to the primary experimental task deployment instruction.

[0057] If the resource container layer in the artificial intelligence experimental cloud platform detects a unified experimental task deployment instruction corresponding to an intermediate experimental task deployment instruction, it obtains the second sub-target image resource, the second GPU resource, and the second data storage volume path corresponding to the intermediate experimental task deployment instruction. The resource container layer then creates a container based on the second sub-target image resource, the second GPU resource, and the second data storage volume path corresponding to the intermediate experimental task deployment instruction.

[0058] If the resource container layer in the artificial intelligence experimental cloud platform detects a high-level experimental task deployment instruction corresponding to a unified experimental task deployment instruction, it obtains the third sub-target image resource, the third GPU resource, and the third data storage volume path corresponding to the high-level experimental task deployment instruction. The resource container layer then creates a container based on the third sub-target image resource, the third GPU resource, and the third data storage volume path corresponding to the high-level experimental task deployment instruction.

[0059] In this embodiment, after the namespace for a class, such as Class A, is created in the cluster management layer of the AI ​​experiment cloud platform, the teacher corresponding to the first-authority user account can deploy containers with corresponding AI frameworks and models based on the needs of the AI ​​experiment task and the knowledge level of the students who need to conduct the AI ​​experiment or test. For example, if the teacher of Class A knows that all students in Class A are beginners, the teacher can use their user terminal to trigger a basic experiment task deployment instruction and send it to the cloud platform interface layer of the AI ​​experiment cloud platform. When the cloud platform interface layer of the AI ​​experiment cloud platform receives and parses the basic experiment task deployment instruction, the resource container layer of the AI ​​experiment cloud platform creates containers according to the first sub-target image resource, first GPU resource, and first data storage volume path corresponding to the basic experiment task deployment instruction. Thus, when students in Class A log into the AI ​​experiment cloud platform and enter their respective containers, they can conduct basic-level AI experiments within those containers. Specifically, a basic level of artificial intelligence experiment requires users to provide partial model code for a specified neural network in a container, and then prompts students to complete the partial model code by manually editing it to obtain the complete model code.

[0060] If the teacher of class A knows that all students in class A are intermediate-level scholars (possessing more AI-related knowledge than beginners, but not yet proficient), then the teacher can use their user terminal to trigger an intermediate-level experimental task deployment instruction and send it to the cloud platform interface layer of the AI ​​experimental cloud platform. When the cloud platform interface layer receives and parses the intermediate-level experimental task deployment instruction, the resource container layer of the AI ​​experimental cloud platform creates containers according to the second sub-target image resource, second GPU resource, and second data storage volume path corresponding to the intermediate-level experimental task deployment instruction. Thus, when students in class A log into the AI ​​experimental cloud platform and enter their respective containers, they can conduct intermediate-level AI experiments within those containers. Specifically, intermediate-level AI experiments require users to manually edit and complete the code for a specified neural network within the container.

[0061] If the teacher of class A knows that all students in class A are advanced-level scholars (possessing more knowledge of artificial intelligence than intermediate-level scholars, reaching a level of mastery), then the teacher can use their user terminal to trigger an advanced experimental task deployment instruction and send it to the cloud platform interface layer of the artificial intelligence experimental cloud platform. When the cloud platform interface layer receives and parses the advanced experimental task deployment instruction, the resource container layer of the artificial intelligence experimental cloud platform creates containers according to the third sub-target image resource, third GPU resource, and third data storage volume path corresponding to the advanced experimental task deployment instruction. Thus, when students in class A log into the artificial intelligence experimental cloud platform and enter their respective containers, they can conduct advanced-level artificial intelligence experiments within those containers. Specifically, advanced-level artificial intelligence experiments require users to perform black-box attacks, adversarial attacks, and other operations on a specified neural network within the container to achieve the experimental objective of cracking the neural network and obtaining its specific parameters.

[0062] In one embodiment, obtaining the first target image resource, GPU resource, and data storage volume path corresponding to the unified experimental task deployment instruction includes:

[0063] If the resource container layer in the AI ​​experimental cloud platform detects a unified experimental task deployment instruction, it obtains the teaching progress information and teacher teaching tag set corresponding to the unified experimental task deployment instruction, and generates the first target image resource, GPU resource and data storage volume path corresponding to the unified experimental task deployment instruction based on the teaching progress information, teacher teaching tag set and preset resource calling strategy.

[0064] In this embodiment, when the resource container layer in the AI ​​experimental cloud platform detects a unified experimental task deployment instruction, the cloud platform interface layer of the AI ​​experimental cloud platform can parse the unified experimental task deployment instruction to determine whether it includes teaching progress information (e.g., learning the first chapter, fifth section of the AA AI course) and teacher teaching tag sets (e.g., tags such as face recognition and convolutional neural networks). If the unified experimental task deployment instruction parses to obtain teaching progress information and teacher teaching tag sets, then container information such as the first target image resource, GPU resource, and data storage volume path required for container creation is generated according to the teaching progress information, teacher teaching tag sets, and a preset resource call strategy. The resource call strategy can be understood as a pre-set mapping table including several teaching progress information, teacher teaching tag sets, and target image resources, GPU resources, and data storage volume paths. Using teaching progress information and teacher teaching tag sets as search conditions, the corresponding target image resources, GPU resources, and data storage volume paths can be retrieved.

[0065] S102. The artificial intelligence experimental cloud platform verifies the access request. If the access request passes the verification, it obtains the target container instance corresponding to the access request.

[0066] In this embodiment, the cloud platform interface layer in the artificial intelligence experimental cloud platform receives and verifies the access request as an API interface. When the access request passes the verification of the cloud platform interface layer, the target container instance corresponding to the access request is obtained and displayed on the display interface of the target user terminal in a web application manner.

[0067] S103, The target user terminal connects to the target container instance in the artificial intelligence experimental cloud platform.

[0068] In this embodiment, after the target user terminal connects to the target container instance in the artificial intelligence experimental cloud platform, the user of the target user terminal can perform artificial intelligence experimental tasks based on the target container instance displayed on the display interface. Specifically, the web application layer of the artificial intelligence experimental cloud platform communicates with the target user terminal to connect the target user terminal to the target container instance in the artificial intelligence experimental cloud platform and display it on the target user terminal.

[0069] S104. The target container instance in the artificial intelligence experimental cloud platform receives the target data sent by the target user terminal, and stores the target operation data obtained by model calculation and the target data in the storage space corresponding to the physical hardware layer in the artificial intelligence experimental cloud platform.

[0070] In this embodiment, after the user of the target user terminal performs task processing for an artificial intelligence experiment based on the target container instance displayed on the display interface, they can feed back the target data corresponding to the target container instance to the artificial intelligence experiment cloud platform. The target data is passed to the model deployed in the target container instance for calculation to obtain target operation data. At this time, the target operation data and the target data need to be stored in the storage space corresponding to the physical hardware layer in the artificial intelligence experiment cloud platform. Since each container created in the resource container layer of the artificial intelligence experiment cloud platform corresponds to the storage space corresponding to the physical hardware layer, and each container may contain files with a specified path that are persistently stored in the storage space of a pre-specified persistent storage server in the physical hardware layer. When the user stops and closes the operation on the target container instance, or exits the target container instance due to a failure, the target data generated by the user's operation on the target container instance before the closing or exit operation and the target operation data obtained by the model calculation are automatically stored in the storage space corresponding to the physical hardware layer of the artificial intelligence experiment cloud platform. When a user re-enters the target container instance, the AI ​​experimental cloud platform automatically retrieves the target operation data from the storage space corresponding to the physical hardware layer to restore the target container instance to its previous state, so that the user can continue to operate on the target container instance upon re-entry.

[0071] Because the operation data of each container is persisted to a dynamically created persistent volume, the AI ​​experimental cloud platform can automatically save historical operation data to the persistent volume, whether the user actively exits the container instance or is passively exited due to a failure. This allows the user to retrieve historical operation data and continue operating on the container instance upon re-entering the container. Furthermore, the AI ​​experimental cloud platform can create one persistent volume declaration (PVC) and one persistent volume (PV) for each container. Even if a user's container instance's Pod (a Pod is the smallest unit that can be created and managed in a Kubernetes system, and is the smallest resource object model created or deployed by the user) is deleted during the filtering process, its persistent volume (PV) will be retained. When the user logs in again, the persistent volume (PV) will be attached to the new Pod to restore the container instance.

[0072] Furthermore, after the target user terminal establishes a communication connection with the Web application layer of the AI ​​experimental cloud platform, it can also implement dynamic proxying of personal container instances based on Redis. Specifically, the cluster management layer of the AI ​​experimental cloud platform monitors changes in container instance creation to obtain the corresponding IP address in real time. This process can be understood as follows: when a container instance is modified by a user using a user terminal, the AI ​​experimental cloud platform can obtain the user terminal's IP address and store it in a Redis cache database along with a key generated according to encryption rules. Moreover, the AI ​​experimental cloud platform uses Nginx as the unified gateway for ordinary users to access container instances, and the Nginx gateway and the Redis cache database exchange data using Lua scripts. When a user logs into the Web application layer using a user terminal to access the container instance interface, encrypted parameters will be automatically included.

[0073] The AI ​​experimental cloud platform also monitors virtual IP changes caused by container instance changes. Through a packaged Kubernetes native API, it promptly writes these virtual IP changes to the Redis cache database, refreshing the IP data in the Redis cache. When a user's access request reaches the Nginx gateway, the gateway analyzes the URL of the request to obtain encrypted parameters and queries the Redis cache database for the corresponding real address. If the real address can be found in the Redis cache database, it is returned and forwarded by the Nginx gateway to the container instance, completing the user's login to their personal container instance.

[0074] S105. The target grading permission user terminal obtains the target operation data, obtains the corresponding target grading data based on the target operation data, and sends the target grading data to the artificial intelligence experimental cloud platform; wherein, the target grading permission user terminal is a user terminal with grading permission in the target group to which the target user terminal belongs among the at least one user terminal.

[0075] In this embodiment, after the user of the target user terminal performs AI experiment task processing and uploads target data based on the target container instance displayed on the display interface, the data can be saved to the storage space corresponding to the physical hardware layer in the AI ​​experiment cloud platform. Furthermore, the target data is processed by the model deployed in the target container instance to obtain target operation data. Subsequently, the teacher corresponding to the user of the target user terminal (e.g., student C) logs into the AI ​​experiment cloud platform using their terminal (i.e., the target grading permission user terminal). They can then obtain the target operation data and the target data, obtain the corresponding target grading data based on the target data and the target operation data, and send the target grading data to the AI ​​experiment cloud platform. This enables revision operations such as grading and annotation of the target operation data, completing further processing of the student's AI experiment task data.

[0076] In one embodiment, step S105 is followed by:

[0077] If the AI ​​experimental cloud platform receives a private image submission file uploaded by any one of the at least one user terminals, it receives an approval instruction sent by the user terminal with the target approval authority corresponding to the private image submission file, performs private image creation on the private image submission file to obtain the current private image, and stores the current private image in the private image repository of the resource container layer of the AI ​​experimental cloud platform.

[0078] In this embodiment, since the initial storage of basic images corresponding to the AI ​​framework in the AI ​​experimental cloud platform is not extensive, users (administrator-privileged users, first-privileged users, and second-privileged users) can upload image data that does not currently exist in the AI ​​experimental cloud platform using their user terminals. This image data, i.e., a private image submission file (such as a Dockerfile), can be designed by the user or downloaded from the internet. After the private image submission file is uploaded to the AI ​​experimental cloud platform, the administrator-privileged user or the first-privileged user executes the script of the private image submission file and reviews it. After passing the review, the current private image corresponding to the image submission file is stored in the private image repository of the resource container layer of the AI ​​experimental cloud platform. Therefore, this approach allows for the expansion of the basic image types in the AI ​​experimental cloud platform.

[0079] This method enables the pre-construction of AI-related basic frameworks and algorithm models within containers of an AI experimental cloud platform. Users can then directly complete AI-related tasks within the container, without needing to upgrade the hardware performance of their local terminals. Furthermore, it avoids manually building the AI-related basic frameworks and algorithm models on the user's local terminal, thus improving the processing efficiency of AI-related tasks.

[0080] This application also provides a data processing system for an AI experimental cloud platform based on Kubernetes (K8S). This K8S-based AI experimental cloud platform data processing system is used to execute any embodiment of the aforementioned K8S-based AI experimental cloud platform data processing method. Specifically, please refer to... Figure 3 , Figure 3 This is a schematic block diagram of the K8S-based artificial intelligence experimental cloud platform data processing system 100 provided in the embodiments of this application.

[0081] Among them, such as Figure 3 As shown, the K8S-based artificial intelligence experimental cloud platform data processing system 100 includes at least one user terminal 101 and an artificial intelligence experimental cloud platform 102.

[0082] The target user terminal is used to send the container access request to the artificial intelligence experimental cloud platform; wherein, the target user terminal is any one of the at least one user terminal 101.

[0083] In this embodiment, when a user logs into the AI ​​experiment cloud platform (which can be considered a container cluster using Kubernetes container orchestration technology or a cloud server cluster) using a user terminal such as a cloud terminal, smartphone, PC (mainly desktop computer), or tablet, the user sends a container access request to the AI ​​experiment cloud platform using the target user terminal. Specifically, the user can connect to the AI ​​experiment cloud platform based on the integrated development environment (such as VS Code or PyCharm) on the target user terminal, thereby further entering the container corresponding to that user to perform AI experiment tasks. Since there is no need to build an AI experiment environment locally on the user terminal, but instead directly logs into the corresponding container in the AI ​​experiment cloud platform to conduct AI-related experiments, the time required to build an AI experiment environment locally is reduced, and it also eliminates the need for users with high levels of professional knowledge to build a local AI experiment environment.

[0084] In one embodiment, the virtual terminal of the Web application layer displayed in the target user terminal of the K8S-based artificial intelligence experimental cloud platform data processing system 100 establishes a communication connection with the target container instance in the artificial intelligence experimental cloud platform through the WebSocket protocol.

[0085] In this embodiment, the Web application layer in the AI ​​experimental cloud platform can be regarded as a virtual terminal simulated by a JavaScript library. This virtual terminal is displayed on the target user terminal. The virtual terminal of the Web application layer displayed on the target user terminal establishes a communication connection with the target container instance in the AI ​​experimental cloud platform through the WebSocket protocol to maintain state and perform front-end and back-end interaction. Therefore, this is an implementation method of dynamically connecting a personal container instance via an AI experimental cloud platform.

[0086] Of course, if a user connects to the AI ​​experimental cloud platform using an integrated development environment (such as VS Code or PyCharm) on the target user's terminal, it needs to determine whether the user has selected a base image with Jupyter components (specifically, Jupyter-Notebook components). If the user has not selected a base image with Jupyter components, the selected base image will be automatically built into an image that supports SSH protocol access (a secure protocol built on the application and transport layers). Specifically, the AI ​​experimental cloud platform backend will automatically configure a Dockerfile script file, using the user's login account and password as the login account and password for the remote SSH service, and mapping the container's port to the host machine's port. Furthermore, if the user uses VS Code to connect to the AI ​​experimental cloud platform, the user can access the corresponding container instance in the AI ​​experimental cloud platform through the open-ssh plugin in VS Code by accessing the host machine's IP address and port.

[0087] In one embodiment, the artificial intelligence experimental cloud platform 102 in the K8S-based artificial intelligence experimental cloud platform data processing system 100 is further configured to obtain a list of user accounts corresponding to the user account creation instruction if a user account creation instruction for an administrator is received.

[0088] The cluster management layer in the artificial intelligence experimental cloud platform creates namespaces corresponding to the user account list.

[0089] In this embodiment, at least three types of user accounts with different permissions are pre-defined in the AI ​​experimental cloud platform: administrator user accounts, first-permission user accounts (such as teacher user accounts), and second-permission user accounts (such as student user accounts). Administrator user accounts have the authority to manage all data within the entire AI experimental cloud platform; for example, creating multiple namespaces in the cluster management layer of the AI ​​experimental cloud platform is one such permission. First-permission user accounts have the authority to create multiple containers within their corresponding namespaces for second-permission user accounts to log in and use. Second-permission user accounts only have the authority to log in to the corresponding containers within their respective namespaces to process AI experimental tasks.

[0090] In this system, an administrator-privileged user account can receive a list of user accounts to be created from a user corresponding to a first-privilege user account. The administrator, after logging into the AI ​​Experiment Cloud Platform, can then combine the teacher's name and the student's class name from the user's bill list to create a combined name. The administrator then creates a namespace within the cluster management layer of the AI ​​Experiment Cloud Platform using this combined name. This can be visualized as the administrator creating a class-specific namespace for the class taught by the teacher whose name corresponds to that teacher's name on the AI ​​Experiment Cloud Platform. Furthermore, the administrator can also match the number of second-privilege user accounts (i.e., the total number of student names in the user's bill list is the same as the total number of second-privilege user accounts) to the student name list (or student ID list) included in the user's bill list.

[0091] Of course, when creating multiple containers within each namespace according to actual needs, it could be that an administrator-privileged user account creates a corresponding number of containers based on the requirements of the AI ​​experiment task and the student name list included in the user billing list; or it could be that a teacher with first-privilege user accounts creates a corresponding number of containers based on the requirements of the AI ​​experiment task and the student name list included in the user billing list. In the AI ​​experiment cloud platform, the relevant information for the namespaces is stored in the cluster management layer, and the containers corresponding to each namespace are deployed in the resource container layer of the AI ​​experiment cloud platform.

[0092] In one embodiment, after the cluster management layer in the artificial intelligence experimental cloud platform creates a namespace corresponding to the user account list, it further includes:

[0093] If the resource container layer in the artificial intelligence experimental cloud platform detects a unified experimental task deployment instruction, it obtains the first target image resource, GPU resource, and data storage volume path corresponding to the unified experimental task deployment instruction. The resource container layer then creates a container based on the first target image resource, GPU resource, and data storage volume path corresponding to the unified experimental task deployment instruction.

[0094] If the resource container layer in the artificial intelligence experimental cloud platform detects a personalized container deployment instruction, it obtains a second target image resource corresponding to the personalized container deployment instruction. The resource container layer then creates a container based on the second target image resource corresponding to the personalized container deployment instruction and the pre-stored data storage volume path.

[0095] In this embodiment, after the namespace for a specific class, such as Class A, is created in the cluster management layer of the artificial intelligence experimental cloud platform, the teacher corresponding to the first-authority user account can generate a unified experimental task deployment instruction based on the requirements of the artificial intelligence experimental task. This unified experimental task deployment instruction specifically sets the first target image resource, GPU resource, and data storage volume path. Subsequently, the resource container layer creates containers corresponding to the first target image resource, GPU resource, and data storage volume path specified in the unified experimental task deployment instruction.

[0096] The first target image resource can be selected from one or more of the following: images integrating frameworks such as TensorFlow, Caffe, and PyTorch required for experiments in the field of artificial intelligence; or images integrating artificial intelligence-related neural network models such as deep neural networks (DNN), convolutional neural networks (CNN), and object detection networks such as YoLoV1 to V5. More specifically, the first target image resource can be selected from the TensorFlow framework, and a convolutional neural network (CNN) can be deployed on the TensorFlow framework.

[0097] Once the teacher corresponding to the first-authority user account generates a unified experimental task deployment instruction based on the requirements of the artificial intelligence experimental task, and completes the creation of multiple containers corresponding to the namespace in the resource container layer, the initial environment setup for the artificial intelligence experimental task is completed.

[0098] Of course, after the namespace for a class, such as Class A, is created in the cluster management layer of the AI ​​experiment cloud platform, the student corresponding to the second-authority user account can generate a personalized container deployment instruction based on their individual needs for AI experiment tasks. This personalized container deployment instruction is then sent by the cloud platform interface layer of the AI ​​experiment cloud platform to the user terminal used by the teacher corresponding to the first-authority user account. Once the teacher approves the personalized container deployment instruction on the user terminal, the resource container layer of the AI ​​experiment cloud platform creates a container based on the second target image resource corresponding to the personalized container deployment instruction and the pre-stored data storage volume path. Similarly, the second target image resource can be selected from one or more of the following: images integrating frameworks such as TensorFlow, Caffe, and PyTorch required for AI-related experiments; or images integrating AI-related neural network models such as Deep Neural Networks (DNN), Convolutional Neural Networks (CNN), and object detection networks such as YoLoV1 to V5. More specifically, the second target image resource can select the TensorFlow framework and deploy the YoLoV5 object detection network on the TensorFlow framework.

[0099] Furthermore, when a student with the second-level user account generates a personalized container deployment instruction based on their individual needs for conducting AI experiments, GPU resources are not allocated by default. This means that containers created based on the personalized deployment instruction are ordinary server containers, not GPU server containers. However, if the personalized deployment instruction requires the use of a GPU server container, the instruction is sent to the user terminal used by the teacher with the first-level user account. Once the teacher approves the personalized deployment instruction on their user terminal, the resource container layer of the AI ​​experiment cloud platform creates a container based on the second target image resources, GPU resources, and pre-stored data storage volume paths corresponding to the personalized deployment instruction.

[0100] In one embodiment, if the resource container layer in the artificial intelligence experimental cloud platform detects a unified experimental task deployment instruction, it obtains the first target image resource, GPU resource, and data storage volume path corresponding to the unified experimental task deployment instruction. The resource container layer then creates a container based on the first target image resource, GPU resource, and data storage volume path corresponding to the unified experimental task deployment instruction, including:

[0101] If the resource container layer in the artificial intelligence experimental cloud platform detects a primary experimental task deployment instruction corresponding to a unified experimental task deployment instruction, it obtains the first sub-target image resource, GPU resource, and data storage volume path corresponding to the primary experimental task deployment instruction. The resource container layer then creates a container based on the first sub-target image resource, first GPU resource, and first data storage volume path corresponding to the primary experimental task deployment instruction.

[0102] If the resource container layer in the artificial intelligence experimental cloud platform detects a unified experimental task deployment instruction corresponding to an intermediate experimental task deployment instruction, it obtains the second sub-target image resource, the second GPU resource, and the second data storage volume path corresponding to the intermediate experimental task deployment instruction. The resource container layer then creates a container based on the second sub-target image resource, the second GPU resource, and the second data storage volume path corresponding to the intermediate experimental task deployment instruction.

[0103] If the resource container layer in the artificial intelligence experimental cloud platform detects a high-level experimental task deployment instruction corresponding to a unified experimental task deployment instruction, it obtains the third sub-target image resource, the third GPU resource, and the third data storage volume path corresponding to the high-level experimental task deployment instruction. The resource container layer then creates a container based on the third sub-target image resource, the third GPU resource, and the third data storage volume path corresponding to the high-level experimental task deployment instruction.

[0104] In this embodiment, after the namespace for a class, such as Class A, is created in the cluster management layer of the AI ​​experiment cloud platform, the teacher corresponding to the first-authority user account can deploy containers with corresponding AI frameworks and models based on the needs of the AI ​​experiment task and the knowledge level of the students who need to conduct the AI ​​experiment or test. For example, if the teacher of Class A knows that all students in Class A are beginners, the teacher can use their user terminal to trigger a basic experiment task deployment instruction and send it to the cloud platform interface layer of the AI ​​experiment cloud platform. When the cloud platform interface layer of the AI ​​experiment cloud platform receives and parses the basic experiment task deployment instruction, the resource container layer of the AI ​​experiment cloud platform creates containers according to the first sub-target image resource, first GPU resource, and first data storage volume path corresponding to the basic experiment task deployment instruction. Thus, when students in Class A log into the AI ​​experiment cloud platform and enter their respective containers, they can conduct basic-level AI experiments within those containers. Specifically, a basic level of artificial intelligence experiment requires users to provide partial model code for a specified neural network in a container, and then prompts students to complete the partial model code by manually editing it to obtain the complete model code.

[0105] If the teacher of class A knows that all students in class A are intermediate-level scholars (possessing more AI-related knowledge than beginners, but not yet proficient), then the teacher can use their user terminal to trigger an intermediate-level experimental task deployment instruction and send it to the cloud platform interface layer of the AI ​​experimental cloud platform. When the cloud platform interface layer receives and parses the intermediate-level experimental task deployment instruction, the resource container layer of the AI ​​experimental cloud platform creates containers according to the second sub-target image resource, second GPU resource, and second data storage volume path corresponding to the intermediate-level experimental task deployment instruction. Thus, when students in class A log into the AI ​​experimental cloud platform and enter their respective containers, they can conduct intermediate-level AI experiments within those containers. Specifically, intermediate-level AI experiments require users to manually edit and complete the code for a specified neural network within the container.

[0106] If the teacher of class A knows that all students in class A are advanced-level scholars (possessing more knowledge of artificial intelligence than intermediate-level scholars, reaching a level of mastery), then the teacher can use their user terminal to trigger an advanced experimental task deployment instruction and send it to the cloud platform interface layer of the artificial intelligence experimental cloud platform. When the cloud platform interface layer receives and parses the advanced experimental task deployment instruction, the resource container layer of the artificial intelligence experimental cloud platform creates containers according to the third sub-target image resource, third GPU resource, and third data storage volume path corresponding to the advanced experimental task deployment instruction. Thus, when students in class A log into the artificial intelligence experimental cloud platform and enter their respective containers, they can conduct advanced-level artificial intelligence experiments within those containers. Specifically, advanced-level artificial intelligence experiments require users to perform black-box attacks, adversarial attacks, and other operations on a specified neural network within the container to achieve the experimental objective of cracking the neural network and obtaining its specific parameters.

[0107] In one embodiment, obtaining the first target image resource, GPU resource, and data storage volume path corresponding to the unified experimental task deployment instruction includes:

[0108] If the resource container layer in the AI ​​experimental cloud platform detects a unified experimental task deployment instruction, it obtains the teaching progress information and teacher teaching tag set corresponding to the unified experimental task deployment instruction, and generates the first target image resource, GPU resource and data storage volume path corresponding to the unified experimental task deployment instruction based on the teaching progress information, teacher teaching tag set and preset resource calling strategy.

[0109] In this embodiment, when the resource container layer in the AI ​​experimental cloud platform detects a unified experimental task deployment instruction, the cloud platform interface layer of the AI ​​experimental cloud platform can parse the unified experimental task deployment instruction to determine whether it includes teaching progress information (e.g., learning the first chapter, fifth section of the AA AI course) and teacher teaching tag sets (e.g., tags such as face recognition and convolutional neural networks). If the unified experimental task deployment instruction parses to obtain teaching progress information and teacher teaching tag sets, then container information such as the first target image resource, GPU resource, and data storage volume path required for container creation is generated according to the teaching progress information, teacher teaching tag sets, and a preset resource call strategy. The resource call strategy can be understood as a pre-set mapping table including several teaching progress information, teacher teaching tag sets, and target image resources, GPU resources, and data storage volume paths. Using teaching progress information and teacher teaching tag sets as search conditions, the corresponding target image resources, GPU resources, and data storage volume paths can be retrieved.

[0110] The artificial intelligence experimental cloud platform 102 is used to verify the access request. If the access request passes the verification, the target container instance corresponding to the access request is obtained.

[0111] In this embodiment, the cloud platform interface layer in the artificial intelligence experimental cloud platform receives and verifies the access request as an API interface. When the access request passes the verification of the cloud platform interface layer, the target container instance corresponding to the access request is obtained and displayed on the display interface of the target user terminal in a web application manner.

[0112] The artificial intelligence experimental cloud platform 102 is also used to connect to the target container instance in the artificial intelligence experimental cloud platform.

[0113] In this embodiment, after the target user terminal connects to the target container instance in the artificial intelligence experimental cloud platform, the user of the target user terminal can perform artificial intelligence experimental tasks based on the target container instance displayed on the display interface. Specifically, the web application layer of the artificial intelligence experimental cloud platform communicates with the target user terminal to connect the target user terminal to the target container instance in the artificial intelligence experimental cloud platform and display it on the target user terminal.

[0114] The artificial intelligence experimental cloud platform 102 is also used to receive target data sent by the target user terminal by the target container instance, and store the target operation data obtained by model calculation of the target data and the target data in the storage space corresponding to the physical hardware layer in the artificial intelligence experimental cloud platform.

[0115] In this embodiment, after the user of the target user terminal performs task processing for an artificial intelligence experiment based on the target container instance displayed on the display interface, they can feed back the target data corresponding to the target container instance to the artificial intelligence experiment cloud platform. The target data is passed to the model deployed in the target container instance for calculation to obtain target operation data. At this time, the target operation data and the target data need to be stored in the storage space corresponding to the physical hardware layer in the artificial intelligence experiment cloud platform. Since each container created in the resource container layer of the artificial intelligence experiment cloud platform corresponds to the storage space corresponding to the physical hardware layer, and each container may contain files with a specified path that are persistently stored in the storage space of a pre-specified persistent storage server in the physical hardware layer. When the user stops and closes the operation on the target container instance, or exits the target container instance due to a failure, the target data generated by the user's operation on the target container instance before the closing or exit operation and the target operation data obtained by the model calculation are automatically stored in the storage space corresponding to the physical hardware layer of the artificial intelligence experiment cloud platform. When a user re-enters the target container instance, the AI ​​experimental cloud platform automatically retrieves the target operation data from the storage space corresponding to the physical hardware layer to restore the target container instance to its previous state, so that the user can continue to operate on the target container instance upon re-entry.

[0116] Because the operation data of each container is persisted to a dynamically created persistent volume, the AI ​​experimental cloud platform can automatically save historical operation data to the persistent volume, whether the user actively exits the container instance or is passively exited due to a failure. This allows the user to retrieve historical operation data and continue operating on the container instance upon re-entering the container. Furthermore, the AI ​​experimental cloud platform can create one persistent volume declaration (PVC) and one persistent volume (PV) for each container. Even if a user's container instance's Pod (a Pod is the smallest unit that can be created and managed in a Kubernetes system, and is the smallest resource object model created or deployed by the user) is deleted during the filtering process, its persistent volume (PV) will be retained. When the user logs in again, the persistent volume (PV) will be attached to the new Pod to restore the container instance.

[0117] Furthermore, after the target user terminal establishes a communication connection with the Web application layer of the AI ​​experimental cloud platform, it can also implement dynamic proxying of personal container instances based on Redis. Specifically, the cluster management layer of the AI ​​experimental cloud platform monitors changes in container instance creation to obtain the corresponding IP address in real time. This process can be understood as follows: when a container instance is modified by a user using a user terminal, the AI ​​experimental cloud platform can obtain the user terminal's IP address and store it in a Redis cache database along with a key generated according to encryption rules. Moreover, the AI ​​experimental cloud platform uses Nginx as the unified gateway for ordinary users to access container instances, and the Nginx gateway and the Redis cache database exchange data using Lua scripts. When a user logs into the Web application layer using a user terminal to access the container instance interface, encrypted parameters will be automatically included.

[0118] The AI ​​experimental cloud platform also monitors virtual IP changes caused by container instance changes. Through a packaged Kubernetes native API, it promptly writes these virtual IP changes to the Redis cache database, refreshing the IP data in the Redis cache. When a user's access request reaches the Nginx gateway, the gateway analyzes the URL of the request to obtain encrypted parameters and queries the Redis cache database for the corresponding real address. If the real address can be found in the Redis cache database, it is returned and forwarded by the Nginx gateway to the container instance, completing the user's login to their personal container instance.

[0119] A target grading permission user terminal is used to acquire the target operation data, obtain corresponding target grading data based on the target operation data, and send the target grading data to the artificial intelligence experimental cloud platform; wherein, the target grading permission user terminal is a user terminal with grading permission in the target group to which the target user terminal belongs among the at least one user terminal.

[0120] In this embodiment, after the user of the target user terminal performs AI experiment task processing and uploads target data based on the target container instance displayed on the display interface, the data can be saved to the storage space corresponding to the physical hardware layer in the AI ​​experiment cloud platform. Furthermore, the target data is processed by the model deployed in the target container instance to obtain target operation data. Subsequently, the teacher corresponding to the user of the target user terminal (e.g., student C) logs into the AI ​​experiment cloud platform using their terminal (i.e., the target grading permission user terminal). They can then obtain the target operation data and the target data, obtain the corresponding target grading data based on the target data and the target operation data, and send the target grading data to the AI ​​experiment cloud platform. This enables revision operations such as grading and annotation of the target operation data, completing further processing of the student's AI experiment task data.

[0121] In one embodiment, the artificial intelligence experimental cloud platform 102 in the K8S-based artificial intelligence experimental cloud platform data processing system 100 is further configured to, upon receiving a private image submission file uploaded by any one of the at least one user terminal, receive an approval instruction corresponding to the private image submission file and sent by the user terminal with the target approval authority, perform private image creation on the private image submission file to obtain a current private image, and store the current private image in the private image repository of the resource container layer of the artificial intelligence experimental cloud platform.

[0122] In this embodiment, since the initial storage of basic images corresponding to the AI ​​framework in the AI ​​experimental cloud platform is not extensive, users (administrator-privileged users, first-privileged users, and second-privileged users) can upload image data that does not currently exist in the AI ​​experimental cloud platform using their user terminals. This image data, i.e., a private image submission file (such as a Dockerfile), can be designed by the user or downloaded from the internet. After the private image submission file is uploaded to the AI ​​experimental cloud platform, the administrator-privileged user or the first-privileged user executes the script of the private image submission file and reviews it. After passing the review, the current private image corresponding to the image submission file is stored in the private image repository of the resource container layer of the AI ​​experimental cloud platform. Therefore, this approach allows for the expansion of the basic image types in the AI ​​experimental cloud platform.

[0123] This method enables the pre-construction of AI-related basic frameworks and algorithm models within containers of an AI experimental cloud platform. Users can then directly complete AI-related tasks within the container, without needing to upgrade the hardware performance of their local terminals. Furthermore, it avoids manually building the AI-related basic frameworks and algorithm models on the user's local terminal, thus improving the processing efficiency of AI-related tasks.

[0124] This system enables the pre-construction of AI-related basic frameworks and algorithm models within containers on an AI experimental cloud platform. Users can directly complete AI-related tasks within the container without needing to upgrade the hardware performance of their local terminals. Furthermore, it eliminates the need for manual construction of AI-related basic frameworks and algorithm models on the user's local terminal, thus improving the processing efficiency of AI-related tasks.

[0125] The aforementioned data processing system for the Kubernetes-based AI experimental cloud platform can be implemented as a computer program, which can be used in various ways, such as... Figure 4 It runs on the computer device shown.

[0126] Please see Figure 4 , Figure 4 This is a schematic block diagram of a computer device provided in an embodiment of this application. The computer device 500 is a server, or it can be a server cluster. The server can be a standalone server, or it can be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms.

[0127] See Figure 4 The computer device 500 includes a processor 502, a memory, and a network interface 505 connected via a device bus 501, wherein the memory may include a storage medium 503 and internal memory 504.

[0128] The storage medium 503 can store the operating system 5031 and the computer program 5032. When the computer program 5032 is executed, it enables the processor 502 to execute the data processing method of the K8S-based artificial intelligence experimental cloud platform.

[0129] The processor 502 provides computing and control capabilities to support the operation of the entire computer device 500.

[0130] The internal memory 504 provides an environment for the computer program 5032 in the storage medium 503 to run. When the computer program 5032 is executed by the processor 502, the processor 502 can execute the data processing method of the K8S-based artificial intelligence experimental cloud platform.

[0131] This network interface 505 is used for network communication, such as providing data transmission. Those skilled in the art will understand that... Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device 500 to which the present application is applied. The specific computer device 500 may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0132] The processor 502 is used to run the computer program 5032 stored in the memory to implement the data processing method of the K8S-based artificial intelligence experimental cloud platform disclosed in the embodiments of this application.

[0133] Those skilled in the art will understand that Figure 4 The embodiments of the computer device shown do not constitute a limitation on the specific configuration of the computer device. In other embodiments, the computer device may include more or fewer components than illustrated, or combine certain components, or have different component arrangements. For example, in some embodiments, the computer device may include only memory and a processor. In such embodiments, the structure and function of the memory and processor are different from those shown. Figure 4 The embodiments shown are consistent and will not be described again here.

[0134] It should be understood that in the embodiments of this application, the processor 502 may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0135] In another embodiment of this application, a computer-readable storage medium is provided. This computer-readable storage medium can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. The computer-readable storage medium stores a computer program, wherein when executed by a processor, the computer program implements the data processing method for the K8S-based artificial intelligence experimental cloud platform disclosed in the embodiments of this application.

[0136] Those skilled in the art will readily understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0137] In the several embodiments provided in this application, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. Units with the same function may be grouped into one unit. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, or it may be an electrical, mechanical, or other form of connection.

[0138] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of the embodiments of this application, depending on actual needs.

[0139] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0140] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, a backend server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), magnetic disks, or optical disks.

[0141] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A data processing method for an artificial intelligence experimental cloud platform based on Kubernetes, applied to an artificial intelligence experimental cloud platform system, characterized in that, The artificial intelligence experimental cloud platform system includes at least one user terminal and an artificial intelligence experimental cloud platform. The artificial intelligence experimental cloud platform includes a web application layer, a cloud platform interface layer, a cluster management layer, a resource container layer, and a physical hardware layer; the method includes: The target user terminal sends a container access request to the artificial intelligence experimental cloud platform; wherein, the target user terminal is any one of the at least one user terminal; The artificial intelligence experimental cloud platform verifies the access request. If the access request passes the verification, it obtains the target container instance corresponding to the access request. The target user terminal is connected to the target container instance in the artificial intelligence experimental cloud platform; The target container instance in the artificial intelligence experimental cloud platform receives the target data sent by the target user terminal, and stores the target operation data obtained by model calculation and the target data in the storage space corresponding to the physical hardware layer in the artificial intelligence experimental cloud platform. The target grading permission user terminal obtains the target operation data, obtains the corresponding target grading data based on the target operation data, and sends the target grading data to the artificial intelligence experimental cloud platform; wherein, the target grading permission user terminal is a user terminal with grading permission in the target group to which the target user terminal belongs among the at least one user terminal; Before the target user terminal sends the container access request to the artificial intelligence experimental cloud platform, the following steps are also included: If the artificial intelligence experimental cloud platform receives a user account creation instruction from a user with administrator privileges, it obtains a list of user accounts corresponding to the user account creation instruction. The cluster management layer in the artificial intelligence experimental cloud platform creates a namespace corresponding to the user account list; After the cluster management layer in the artificial intelligence experimental cloud platform creates namespaces corresponding to the user account list, it also includes: If the resource container layer in the artificial intelligence experimental cloud platform detects a unified experimental task deployment instruction, it obtains the first target image resource, GPU resource, and data storage volume path corresponding to the unified experimental task deployment instruction. The resource container layer then creates a container based on the first target image resource, GPU resource, and data storage volume path corresponding to the unified experimental task deployment instruction. If the resource container layer in the artificial intelligence experimental cloud platform detects a personalized container deployment instruction, it obtains the second target image resource corresponding to the personalized container deployment instruction. The resource container layer then creates a container based on the second target image resource corresponding to the personalized container deployment instruction and the pre-stored data storage volume path. The first target image resource integrates framework images of TensorFlow, Caffe and PyTorch required for experiments in the field of artificial intelligence, or integrates any one or more of deep neural networks, convolutional neural networks and object detection networks. The personalized container deployment instruction is sent from the cloud platform interface layer of the artificial intelligence experimental cloud platform to the user terminal used by the teacher corresponding to the first-authority user account. After the teacher corresponding to the first-authority user account approves the personalized container deployment instruction on the user terminal, the resource container layer of the artificial intelligence experimental cloud platform executes the step of obtaining the second target image resource corresponding to the personalized container deployment instruction, and the resource container layer creates a container according to the second target image resource corresponding to the personalized container deployment instruction and the pre-stored data storage volume path.

2. The data processing method for the K8S-based artificial intelligence experimental cloud platform according to claim 1, characterized in that, The target user terminal is connected to the target container instance in the artificial intelligence experimental cloud platform, including: The virtual terminal of the Web application layer displayed in the target user terminal establishes a communication connection with the target container instance in the artificial intelligence experimental cloud platform through the WebSocket protocol.

3. The data processing method for the K8S-based artificial intelligence experimental cloud platform according to claim 1, characterized in that, If the resource container layer in the AI ​​experimental cloud platform detects a unified experimental task deployment instruction, it obtains the first target image resource, GPU resource, and data storage volume path corresponding to the unified experimental task deployment instruction. The resource container layer then creates a container based on the first target image resource, GPU resource, and data storage volume path corresponding to the unified experimental task deployment instruction, including: If the resource container layer in the artificial intelligence experimental cloud platform detects a primary experimental task deployment instruction corresponding to a unified experimental task deployment instruction, it obtains the first sub-target image resource, GPU resource, and data storage volume path corresponding to the primary experimental task deployment instruction. The resource container layer then creates a container based on the first sub-target image resource, first GPU resource, and first data storage volume path corresponding to the primary experimental task deployment instruction. If the resource container layer in the artificial intelligence experimental cloud platform detects a unified experimental task deployment instruction corresponding to an intermediate experimental task deployment instruction, it obtains the second sub-target image resource, the second GPU resource, and the second data storage volume path corresponding to the intermediate experimental task deployment instruction. The resource container layer then creates a container based on the second sub-target image resource, the second GPU resource, and the second data storage volume path corresponding to the intermediate experimental task deployment instruction. If the resource container layer in the artificial intelligence experimental cloud platform detects a high-level experimental task deployment instruction corresponding to a unified experimental task deployment instruction, it obtains the third sub-target image resource, the third GPU resource, and the third data storage volume path corresponding to the high-level experimental task deployment instruction. The resource container layer then creates a container based on the third sub-target image resource, the third GPU resource, and the third data storage volume path corresponding to the high-level experimental task deployment instruction.

4. The data processing method for the K8S-based artificial intelligence experimental cloud platform according to claim 3, characterized in that, The step of obtaining the first target image resource, GPU resource, and data storage volume path corresponding to the unified experimental task deployment instruction includes: If the resource container layer in the AI ​​experimental cloud platform detects a unified experimental task deployment instruction, it obtains the teaching progress information and teacher teaching tag set corresponding to the unified experimental task deployment instruction, and generates the first target image resource, GPU resource and data storage volume path corresponding to the unified experimental task deployment instruction based on the teaching progress information, teacher teaching tag set and preset resource calling strategy.

5. The data processing method for an artificial intelligence experimental cloud platform based on K8S according to any one of claims 1 to 4, characterized in that, After the target grading permission user terminal obtains the target operation data, obtains the corresponding target grading data based on the target operation data, and sends the target grading data to the artificial intelligence experimental cloud platform, the process further includes: If the AI ​​experimental cloud platform receives a private image submission file uploaded by any one of the at least one user terminals, it receives an approval instruction sent by the user terminal with the target approval authority corresponding to the private image submission file, performs private image creation on the private image submission file to obtain the current private image, and stores the current private image in the private image repository of the resource container layer of the AI ​​experimental cloud platform.

6. A data processing system for an artificial intelligence experimental cloud platform based on Kubernetes, characterized in that, It includes at least one user terminal and an artificial intelligence experimental cloud platform; the artificial intelligence experimental cloud platform includes a web application layer, a cloud platform interface layer, a cluster management layer, a resource container layer, and a physical hardware layer; The target user terminal is used to send the container access request to the artificial intelligence experimental cloud platform; wherein the target user terminal is any one of the at least one user terminal. The artificial intelligence experimental cloud platform is used to verify the access request. If the access request passes the verification, the target container instance corresponding to the access request is obtained. The target user terminal is also used to connect to the target container instance in the artificial intelligence experimental cloud platform; The artificial intelligence experimental cloud platform is also used to receive target data sent by the target user terminal through the target container instance, and store the target operation data obtained by model calculation of the target data and the target data in the storage space corresponding to the physical hardware layer in the artificial intelligence experimental cloud platform; A target grading permission user terminal is used to acquire the target operation data, obtain corresponding target grading data based on the target operation data, and send the target grading data to the artificial intelligence experimental cloud platform; wherein, the target grading permission user terminal is a user terminal with grading permission in the target group to which the target user terminal belongs among the at least one user terminal; The artificial intelligence experimental cloud platform is also used to obtain a list of user accounts corresponding to the user account creation instruction if it receives a user account creation instruction from a user with administrator privileges. The cluster management layer in the artificial intelligence experimental cloud platform creates a namespace corresponding to the user account list; After the cluster management layer in the artificial intelligence experimental cloud platform creates namespaces corresponding to the user account list, it also includes: If the resource container layer in the artificial intelligence experimental cloud platform detects a unified experimental task deployment instruction, it obtains the first target image resource, GPU resource, and data storage volume path corresponding to the unified experimental task deployment instruction. The resource container layer then creates a container based on the first target image resource, GPU resource, and data storage volume path corresponding to the unified experimental task deployment instruction. If the resource container layer in the artificial intelligence experimental cloud platform detects a personalized container deployment instruction, it obtains the second target image resource corresponding to the personalized container deployment instruction. The resource container layer then creates a container based on the second target image resource corresponding to the personalized container deployment instruction and the pre-stored data storage volume path. The first target image resource integrates framework images of TensorFlow, Caffe and PyTorch required for experiments in the field of artificial intelligence, or integrates any one or more of deep neural networks, convolutional neural networks and object detection networks. The personalized container deployment instruction is sent from the cloud platform interface layer of the artificial intelligence experimental cloud platform to the user terminal used by the teacher corresponding to the first-authority user account. After the teacher corresponding to the first-authority user account approves the personalized container deployment instruction on the user terminal, the resource container layer of the artificial intelligence experimental cloud platform performs the operation of obtaining the second target image resource corresponding to the personalized container deployment instruction. The resource container layer creates a container according to the second target image resource corresponding to the personalized container deployment instruction and the pre-stored data storage volume path.

7. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the data processing method for the K8S-based artificial intelligence experimental cloud platform as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, causes the processor to perform the data processing method for the K8S-based artificial intelligence experimental cloud platform as described in any one of claims 1 to 5.

Images