VNPU computing power scheduling method oriented to teaching practical training, terminal equipment and storage medium
By virtualizing the physical NPU into a vNPU computing power pool and combining it with containerized scheduling, the problems of resource contention and poor stability in traditional teaching and training are solved. This achieves efficient resource reuse and multi-user isolation, improving NPU resource utilization and training flexibility.
Patent Information
- Application Number
- CN202511401867.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-09-28
- Publication Date
- 2025-11-28
AI Technical Summary
In traditional teaching and training, the physical NPU resources are used in a dedicated manner for the entire card, which makes it difficult to meet the needs of multiple users sharing the physical NPU. This can easily lead to problems such as resource contention and inconsistent environment configuration, resulting in low resource utilization and poor training stability.
By virtualizing the physical NPU to form a vNPU computing power pool, and combining containerized scheduling to dynamically allocate independent logical computing power units to student terminals, the virtualization layer is established using Ascend Docker Runtime for binding, the resource usage status is monitored in real time and dynamically adjusted, and the vNPU computing power parameters are accurately matched according to the type of training task.
It achieves efficient reuse of physical resources and multi-user isolation, supports multiple students to simultaneously access interactive computing environments in large-scale teaching and training, and improves the utilization rate of NPU resources and the flexibility and stability of teaching and training.
Smart Images
Figure CN121029418A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application belongs to the field of data processing, and particularly relates to a vNPU computing power scheduling method for teaching and training, a terminal device and a storage medium. BACKGROUND
[0002] In traditional teaching and training, physical NPU resources are difficult to meet the needs of concurrent training of multiple students due to the whole-card exclusive allocation mode. In traditional technologies, resource competition conflicts and inconsistent environment configurations are prone to occur when multiple users share physical NPUs, resulting in low resource utilization and poor training stability. A new technical means is needed to solve the above technical problems. SUMMARY
[0003] In view of this, the embodiments of the present application provide a vNPU computing power scheduling method for teaching and training, a terminal device and a storage medium, which can solve the problems of low resource utilization and poor training stability in related technologies.
[0004] The first aspect of the present application provides a vNPU computing power scheduling method for teaching and training, comprising: generating a vNPU computing power pool, and receiving a container creation request from a student terminal, wherein the vNPU computing power pool is composed of independent logical computing units divided by virtualization of multiple physical NPUs, and the container creation request is used to request to create a computing power container containing an interactive computing environment; creating the computing power container according to the container creation request, and scheduling a target vNPU from the vNPU computing power pool according to the container creation request; binding the target vNPU to the computing power container; the scheduling device starts the computing power container, and the computing power container provides teaching and training services to the student terminal through the interactive computing environment.
[0005] Optionally, in the first implementation manner of the first aspect of the present application, the step of binding the target vNPU to the to-be-created computing power container comprises: establishing a virtualization layer based on Ascend Docker Runtime; associating and configuring the computing power resources of the target vNPU with the running environment of the computing power container through the virtualization layer to complete the binding, wherein the computing power container exclusively has the computing power access authority of the target vNPU.
[0006] Optionally, in the second implementation manner of the first aspect of the present application, after the step of establishing a virtualization layer based on Ascend Docker Runtime, the method further comprises Monitoring a resource usage state of the computing power container in real time; When the resource usage state reaches a preset threshold, adjusting, by the virtualization layer, computing power allocation of the target vNPU to the computing power container.
[0007] Optionally, in a third implementation manner of the first aspect, the step of scheduling a target vNPU from the vNPU computing power pool according to the container creation request comprises: determining a required vNPU computing power parameter according to a type of the real training task carried in the container creation request; matching and scheduling the target vNPU satisfying the vNPU computing power parameter from the vNPU computing power pool according to the vNPU computing power parameter.
[0008] Optionally, in a fourth implementation manner of the first aspect, after the step of scheduling a target vNPU from the vNPU computing power pool according to the container creation request, the method further comprises: monitoring a resource usage state of the computing power container; When the resource usage state satisfies a preset adjustment condition, dynamically adjusting computing power allocation of the target vNPU.
[0009] Optionally, in a fifth implementation manner of the first aspect, the computing power container contains a preset computing framework image, and the computing framework image is used to provide a basic running environment of the interactive computing environment.
[0010] Optionally, in a sixth implementation manner of the first aspect, the computing framework image contains an operating system, a driver and a computing library required by artificial intelligence training.
[0011] Optionally, in a seventh implementation manner of the first aspect, the interactive computing environment is Jupyter-lab, the Jupyter-lab provides a visual interactive interface to the student terminal through a B / S architecture, and the visual interactive interface supports code running, file management and command line operation.
[0012] In a second aspect, an embodiment of the present application provides a terminal device, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, and the processor implements the steps of the vNPU computing power scheduling method for teaching and real training when executing the computer program.
[0013] In a third aspect, an embodiment of the present application provides a computer readable storage medium, which stores a computer program, and the computer program implements the steps of the vNPU computing power scheduling method for teaching and real training when executed by a processor.
[0014] In a fourth aspect, an embodiment of the present application provides a computer program product, which, when running on a terminal device, causes the terminal device to execute the above-mentioned vNPU computing power scheduling method for teaching practice.
[0015] Compared with the prior art, the embodiment of the present application has the beneficial effects that: by virtualizing the physical NPU to form a vNPU computing power pool, and dynamically allocating independent logical computing power units to student terminals in combination with container scheduling, efficient reuse and multi-user isolation of physical resources are realized, which can support the demand of multiple students to simultaneously obtain an interactive computing environment in large-scale teaching practice, and can also ensure independent running of each practice task through the computing power container, thereby improving the NPU resource utilization rate and the flexibility of teaching practice. BRIEF DESCRIPTION OF DRAWINGS
[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor.
[0017] Figure 1 An embodiment of the vNPU computing power scheduling method for teaching practice in the embodiment of the present application is shown in the figure. Figure 2 An embodiment of step S103 of the vNPU computing power scheduling method for teaching practice in the embodiment of the present application is shown in the figure. Figure 3 An embodiment of the vNPU computing power scheduling method for teaching practice in the embodiment of the present application is shown in the figure. Figure 4 An embodiment of step S102 of the vNPU computing power scheduling method for teaching practice in the embodiment of the present application is shown in the figure. Figure 5 An embodiment of the terminal device in the embodiment of the present application is shown in the figure. DETAILED DESCRIPTION
[0018] In order to make the purpose, technical solutions and advantages of the present application more clear, the following will further describe the present application in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative labor are within the protection scope of the present application.
[0019] It should be noted that the terms "comprising", "including", and "having" and any variations thereof in the present specification and claims and the above-described accompanying drawings are intended to cover a non-exclusive inclusion. For example, a process, method, terminal, product or device including a series of steps or units is not limited to the listed steps or units, but optionally further includes steps or units not listed, or optionally further includes other steps or units inherent to the process, method, product or device. In the claims, specification and drawings of the present application, the terms such as "first" and "second" and the like relationship terms are only used to distinguish one entity / operation / object from another entity / operation / object, and do not necessarily require or imply any such real-time relationship or sequence between the entities / operations / objects.
[0020] Reference herein to "embodiment" means that the particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the application. The appearance of the phrase in various places in the specification does not necessarily all refer to the same embodiment, nor is it necessary that a separate or alternative embodiment be developed for each different combination of features described. It is explicitly and implicitly recognized that one or more features of the embodiments described can be combined with one or more other features that are described in the application.
[0021] In traditional teaching practice, physical NPU resources are difficult to meet the needs of concurrent training of multiple students due to the use of whole card exclusive allocation mode, and in traditional technology, resource competition conflicts and inconsistent environment configurations are prone to occur when multiple users share physical NPU, resulting in low resource utilization and poor training stability. A new technical means is needed to solve the above technical problems.
[0022] In view of this, the embodiment of the application provides a vNPU computing power scheduling method, a terminal device and a storage medium for teaching practice, which forms a vNPU computing power pool by virtualizing physical NPU, and dynamically allocates independent logical computing power units to student terminals in combination with containerization scheduling, thereby realizing efficient reuse and multi-user isolation of physical resources, supporting the needs of multiple students to simultaneously obtain interactive computing environments in large-scale teaching practice, and guaranteeing independent operation of each training task through computing power containers, thereby improving NPU resource utilization and flexibility of teaching practice.
[0023] In order to illustrate the technical solutions of the present application, the following will be described through specific embodiments.
[0024] Figure 1A flowchart of a vNPU computing power scheduling method for teaching and training provided by an embodiment of the present application is shown. The method can be applied to a terminal device. The terminal device can be a mobile phone, a tablet computer, a notebook computer, an ultra-mobile personal computer (UMPC), a netbook, etc.
[0025] Specifically, the vNPU computing power scheduling method for teaching and training can include the following steps S101 to S103.
[0026] In step S101, a vNPU computing power pool is generated, and a container creation request from a student terminal is received. The vNPU computing power pool is composed of independent logical computing units divided by virtualization from a plurality of physical NPUs. The container creation request is used to request the creation of a computing power container containing an interactive computing environment.
[0027] In an embodiment of the present application, the terminal device generates a vNPU computing power pool, which is divided into a plurality of independent logical computing units by virtualization technology from a plurality of physical NPUs. At the same time, the terminal device receives a container creation request sent from a student terminal. The container creation request is used to request the creation of a computing power container containing an interactive computing environment.
[0028] Optionally, the interactive computing environment is Jupyter-lab, which provides a visual interactive interface to the student terminal through a B / S architecture. The visual interactive interface supports code running, file management, and command line operation. Using Jupyter-lab as the interactive computing environment and providing a visual interface supporting code running, file management, and command line operation through a B / S architecture improves the interactive convenience of teaching and training and solves the problem of complex operation and slow student adaptation in traditional training.
[0029] In step S102, a computing power container is created according to the container creation request, and a target vNPU is scheduled from the vNPU computing power pool according to the container creation request.
[0030] In an embodiment of the present application, the terminal device creates a corresponding computing power container according to the received container creation request. At the same time, the terminal device schedules an adapted target vNPU from the generated vNPU computing power pool according to the specific requirements of the container creation request.
[0031] Optionally, according to the differences in the training scenarios embodied in the container creation request, vNPUs meeting the basic computing power requirements are preliminarily selected as the candidate range of the target vNPU.
[0032] In step S103, the target vNPU is bound to the computing power container.
[0033] In the embodiment of the present application, the target vNPU obtained by scheduling is bound to the created computing power container, and the computing power association between the two is established, so that the computing power container can stably call the computing power resources of the target vNPU.
[0034] Step S104, the scheduling device starts the computing power container, and the computing power container provides teaching and training services to the student terminal through the interactive computing environment.
[0035] In the embodiment of the present application, the computing power container is started, and after the computing power container is started, teaching and training services are provided to the student terminal sending the request through the interactive computing environment contained in the computing power container, so that the student terminal can carry out training operation based on the environment.
[0036] Optionally, the computing power container contains a preset computing framework image, and the computing framework image is used to provide a basic running environment of the interactive computing environment. By integrating the preset computing framework image in the computing power container to provide the basic running environment of the interactive computing environment, the tedious process of manually building the environment by the student in the teaching and training is avoided, and the inconsistent progress problem caused by environment configuration difference in the traditional training is solved.
[0037] Optionally, the computing framework image contains an operating system, a driver and a computing library required for artificial intelligence training. Wherein, the computing framework image contains the operating system, the driver and the computing library required for artificial intelligence training, which ensures the integrity and compatibility of the interactive computing environment, and solves the experiment failure problem caused by missing components or version mismatch in the traditional training.
[0038] The beneficial effects of the embodiment of the present application compared with the prior art are: by virtualizing the physical NPU to form a vNPU computing power pool, and dynamically allocating independent logical computing power units to student terminals through container scheduling, efficient reuse and multi-user isolation of physical resources are realized, which can support the demand of multiple students simultaneously obtaining interactive computing environment in large-scale teaching and training, and can ensure independent running of each training task through the computing power container, thereby improving the NPU resource utilization and the flexibility of teaching and training.
[0039] The traditional vNPU and container binding mode lacks a dedicated virtualization layer support, and cannot realize strict exclusive access control of the container to the vNPU, which leads to problems such as resource contention and calling conflict in multi-user training, and destroys the stability of the training environment. Based on this, an optional embodiment of the present application is proposed. Referring to Figure 2 , Figure 2 For a specific embodiment of step S103 of the vNPU computing power scheduling method for teaching and training in the embodiment of the present application, step S103 further includes the following specific implementation.
[0040] Step S1031, establishing a virtualization layer based on the Ascend Docker Runtime.
[0041] In an embodiment of the present application, the virtualization layer is established based on the Ascend Docker Runtime, which serves as an intermediate carrier for bridging the target vNPU computing power resources and the computing power container running environment.
[0042] Step S1032, associating and configuring the computing power resources of the target vNPU with the running environment of the computing power container through the virtualization layer to complete the binding, wherein the computing power container exclusively has the access permission to the computing power of the target vNPU.
[0043] In an embodiment of the present application, the computing power resources of the target vNPU are associated and configured with the running environment of the computing power container through the virtualization layer to complete the binding of the target vNPU and the computing power container; wherein, in the binding process, it is determined that the computing power container exclusively has the access permission to the computing power of the target vNPU, preventing illegal calls to the vNPU resources by other containers.
[0044] In an embodiment of the present application, the virtualization layer is established based on the Ascend Docker Runtime to achieve exclusive access binding of the computing power container to the vNPU, effectively enhancing the association stability and isolation of the computing power resources and the container environment, and avoiding conflicts in computing power permissions in multi-user sharing scenarios.
[0045] The traditional vNPU computing power allocation method is fixed allocation, which cannot be flexibly adjusted according to the real-time resource usage state of the container in teaching and training, resulting in problems such as waste of excess computing power or insufficient computing power affecting the progress of training when the computing power demand fluctuates in training tasks. Based on this, an optional embodiment is proposed. Referring to Figure 3 , Figure 3 For a specific embodiment of the step S1031 of the vNPU computing power scheduling method for teaching and training in an embodiment of the present application, the step S1031 further includes the following specific embodiments.
[0046] Step S1033, real-time monitoring of the resource usage state of the computing power container.
[0047] In an embodiment of the present application, after the virtualization layer is established based on the Ascend Docker Runtime, the resource usage state of the computing power container associated with the virtualization layer is monitored in real time to obtain dynamic information such as computing power consumption and resource occupation of the container during the training process.
[0048] Step S1034, when the resource usage state reaches a preset threshold, adjusting the computing power allocation of the target vNPU to the computing power container through the virtualization layer.
[0049] In an embodiment of the present application, the monitored resource usage state of the computing power container is compared with a preset threshold, and when the resource usage state reaches the preset threshold, the computing power allocated to the computing power container by the established virtualization layer is adjusted for the target vNPU.
[0050] Optionally, the preset threshold can be pre-set or dynamically updated according to factors such as the complexity of the training task and the required computing power scale.
[0051] In an embodiment of the present application, by monitoring the resource usage state of the computing power container in real time and adjusting the vNPU computing power allocation with the help of the virtualization layer when the preset threshold is reached, dynamic adaptation of computing power resources can be achieved, and waste or insufficient computing power caused by fixed allocation can be avoided.
[0052] The traditional vNPU computing power scheduling does not consider the differentiated demand of different task types in teaching and training for computing power, and usually adopts a unified allocation mode, which leads to waste of excessive computing power occupied by part of simple tasks, or complex tasks cannot be completed smoothly due to insufficient computing power, affecting the training efficiency and resource utilization. Based on this, an optional embodiment of the present application is proposed. Referring to Figure 4 , Figure 4 For a specific embodiment of step S102 of the vNPU computing power scheduling method for teaching and training in an embodiment of the present application, step S102 further includes the following specific implementation.
[0053] Step S1021, according to the training task type carried in the container creation request, determine the required vNPU computing power parameter.
[0054] In an embodiment of the present application, after receiving the container creation request from the student terminal, the request is parsed to obtain the training task type carried therein, which reflects the specific content of the teaching and training required by the student (such as image recognition training, natural language processing experiment, etc.).
[0055] Optionally, the training task type is identified by a preset task type identification rule (such as a specific keyword in the request field).
[0056] According to the identified training task type, determine the vNPU computing power parameter required to complete the task, which can include computing power size, computing precision support, parallel processing capability, and other key indicators adapted to the task.
[0057] Step S1022, according to the vNPU computing power parameter, match and schedule the target vNPU that meets the vNPU computing power parameter from the vNPU computing power pool.
[0058] In an embodiment of the present application, based on the determined vNPU computing power parameter, each logical computing power unit in the vNPU computing power pool is screened, and the vNPU meeting the parameter requirement is matched and scheduled as the target vNPU to support the corresponding practical training task.
[0059] In an embodiment of the present application, by determining the required vNPU computing power parameter according to the practical training task type and accurately matching and scheduling, the vNPU computing power resource is adaptively matched with the practical training task demand, and the problem of mismatch between computing power and task caused by blind allocation is avoided.
[0060] The traditional vNPU computing power scheduling cannot be adjusted according to the real-time resource usage state of the computing power container during the practical training process after completing the initial allocation, so that when the practical training task load fluctuates, the problems of computing power waste or insufficient computing power affecting task progress are prone to occur. Based on this, an optional embodiment of the present application is proposed. After step S102, the following specific embodiments are further included.
[0061] Step S201, monitoring the resource usage state of the computing power container.
[0062] In an embodiment of the present application, after scheduling the target vNPU from the vNPU computing power pool according to the container creation request, the resource usage state of the computing power container bound to the target vNPU is continuously monitored, and the resource usage state can include real-time information such as computing power occupancy rate, task processing progress, resource load, etc.
[0063] Optionally, the resource usage state is obtained through a preset monitoring frequency (such as once per second) or triggered monitoring (such as when the container performs a specific operation).
[0064] Step S202, when the resource usage state meets the preset adjustment condition, dynamically adjusting the computing power allocation of the target vNPU.
[0065] In an embodiment of the present application, the monitored resource usage state is compared with the preset adjustment condition, and when the resource usage state meets the preset adjustment condition (for example, the computing power occupancy rate is continuously higher than 90% or lower than 10%), the computing power allocated to the computing power container by the target vNPU is dynamically adjusted to adapt to the current actual demand of the container.
[0066] In an embodiment of the present application, by continuously monitoring the resource usage state of the computing power container after scheduling the target vNPU and dynamically adjusting the computing power allocation when the preset condition is met, the vNPU computing power supply can be accurately matched with the real-time demand of the container, and the problems of excess or insufficient computing power caused by fixed allocation are avoided.
[0067] For example, Figure 5As shown, it is a schematic diagram of a terminal device provided by an embodiment of the present application. The terminal device 500 can include a processor 501, a memory 502, and a computer program 503 stored in the memory 502 and executable on the processor 501, for example, a vNPU computing power scheduling program for teaching practice. The processor 501 implements the steps in each of the above vNPU computing power scheduling embodiments for teaching practice when executing the computer program 503.
[0068] The computer program can be divided into one or more modules / units, one or more modules / units are stored in the memory 502 and executed by the processor 501 to complete the present application. One or more modules / units can be a series of computer program instruction segments capable of completing a specific function, which is used to describe the execution process of the computer program in the terminal device.
[0069] The terminal device can include, but is not limited to, the processor 501 and the memory 502. Those skilled in the art can understand that the terminal device can include more or fewer components than those shown, or combine certain components, or different components, for example, the terminal device can also include an input / output device, a network access device, a bus, etc. Figure 5 The terminal device is only an example and does not constitute a limitation on the terminal device, and can include more or fewer components than those shown, or combine certain components, or different components, for example, the terminal device can also include an input / output device, a network access device, a bus, etc.
[0070] The processor 501 can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor.
[0071] The memory 502 can be an internal storage unit of the terminal device, for example, a hard disk or a memory of the terminal device. The memory 502 can also be an external storage device of the terminal device, for example, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the terminal device. Further, the memory 502 can include both the internal storage unit and the external storage device of the terminal device. The memory 502 is used to store computer programs and other programs and data required by the terminal device. The memory 502 can also be used to temporarily store data that has been output or will be output.
[0072] It should be noted that, for the convenience and brevity of description, the structure of the terminal device described above can also refer to the specific description of the structure in the method embodiments, which will not be described here.
[0073] The embodiment of the present application also provides a computer readable storage medium, the computer readable storage medium stores a computer program, and the computer program is executed by a processor to realize the steps in the vNPU computing power scheduling method for teaching practical training.
[0074] The embodiment of the present application provides a computer program product, when the computer program product runs on a mobile terminal, so that the mobile terminal executes the steps in the vNPU computing power scheduling method for teaching practical training.
[0075] In the above embodiments, the description of each embodiment has its own emphasis, and the parts not described or recorded in a certain embodiment can be referred to the related description of other embodiments.
[0076] Those skilled in the art can appreciate that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.
[0077] In the embodiments provided by the present application, it should be understood that the disclosed terminal device and method can be implemented by other ways. For example, the terminal device embodiments described above are only schematic. In addition, the mutual coupling or direct coupling or communication connection between the shown or discussed mutually can be indirect coupling or communication connection through some interfaces, devices or units, and can be electrical, mechanical or other forms.
[0078] The units described as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, i.e. may be located in one place, or may be distributed on multiple network units. Part or all of the units may be selected according to actual needs to achieve the purpose of the embodiment.
[0079] In addition, each functional unit in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.
[0080] The integrated module / unit, if realized in the form of a software functional unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on this understanding, all or part of the processes in the above-mentioned embodiment methods can also be completed by a computer program instructing related hardware, and the computer program can be stored in a computer readable storage medium. When the processor executes the computer program, the steps of each method embodiment can be implemented. The computer program includes computer program code, which can be in the form of source code, object code, executable file or some intermediate form, etc. The computer readable medium can include any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc.
[0081] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than limit them. Although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that the technical solutions recorded in the foregoing embodiments can be modified, or some technical features can be replaced by equivalent ones. These modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.
Claims
1. A vNPU computing power scheduling method for teaching and practical training, characterized in that, include: Generate a vNPU computing power pool and receive a container creation request from a student terminal. The vNPU computing power pool consists of independent logical computing power units formed by virtualization of multiple physical NPUs. The container creation request is used to request the creation of a computing power container containing an interactive computing environment. Based on the container creation request, the computing power container is created, and the target vNPU is scheduled from the vNPU computing power pool based on the container creation request. Bind the target vNPU to the computing power container; The scheduling device activates the computing power container, which then provides teaching and training services to the student terminal through the interactive computing environment.
2. The vNPU computing power scheduling method for teaching and practical training as described in claim 1, characterized in that, The step of binding the target vNPU to the computing power container to be created includes: A virtualization layer is built based on Ascend Docker Runtime; The virtualization layer associates the computing resources of the target vNPU with the runtime environment of the computing container to complete the binding, wherein the computing container exclusively has access to the computing power of the target vNPU.
3. The vNPU computing power scheduling method for teaching and practical training as described in claim 2, characterized in that, Following the step of establishing a virtualization layer based on Ascend Docker Runtime, the method further includes... Real-time monitoring of the resource usage status of the computing power container; When the resource usage status reaches a preset threshold, the virtualization layer adjusts the computing power allocation of the target vNPU to the computing power container.
4. The vNPU computing power scheduling method for teaching and practical training as described in claim 1, characterized in that, The step of scheduling the target vNPU from the vNPU computing power pool according to the container creation request includes: Determine the required vNPU computing power parameters based on the training task type carried in the container creation request; Based on the vNPU computing power parameters, the target vNPU that meets the vNPU computing power parameters is matched and scheduled from the vNPU computing power pool.
5. The vNPU computing power scheduling method for teaching and practical training as described in claim 1, characterized in that, After the step of scheduling the target vNPU from the vNPU computing power pool according to the container creation request, the method further includes: Monitor the resource usage status of the computing power container; When the resource usage status meets the preset adjustment conditions, the computing power allocation of the target vNPU is dynamically adjusted.
6. The vNPU computing power scheduling method for teaching and practical training as described in claim 1, characterized in that, The computing container contains a preset computing framework image, which is used to provide the basic operating environment for the interactive computing environment.
7. The vNPU computing power scheduling method for teaching and practical training as described in claim 6, characterized in that, The computing framework integrates the operating system, drivers, and computing libraries required for artificial intelligence training.
8. The vNPU computing power scheduling method for teaching and practical training as described in claim 1, characterized in that, The interactive computing environment is Jupyter-lab, which provides a visual interactive interface to the student terminal through a B / S architecture. The visual interactive interface supports code execution, file management, and command line operations.
9. A terminal 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 steps of the vNPU computing power scheduling method for teaching and training as described in any one of claims 1 to 8.
10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the vNPU computing power scheduling method for teaching and training as described in any one of claims 1 to 8.