Keyword-based model deployment method, device, equipment and storage medium

Abstract

The application discloses a keyword-based model deployment method and device, equipment and a storage medium. The method comprises the following steps: obtaining a model structure keyword and a model resource keyword; generating a target processing model based on the model structure keyword; the target processing model comprises a plurality of model components; and performing resource configuration on the plurality of model components in the target processing model based on the model resource keyword. The method can simplify the deployment process of a large model service and reduce the difficulty of resource configuration in the large model deployment process.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer, in particular to a model deployment method and device based on keywords, equipment and storage medium. BACKGROUND

[0002] A large model refers to a machine learning model with large-scale parameters and complex computing structure. In order to enable users to easily integrate the powerful capabilities of large models into applications, developers need to deploy models into large model services in the form of interfaces. Therefore, large model service deployment has become an indispensable part of the entire production cycle of large models. However, in the prior art, users usually manually deploy multiple model components into a large model service, and the large model service deployment process is cumbersome. SUMMARY

[0003] The embodiments of the present application provide a model deployment method and device based on keywords, equipment and storage medium, which can simplify the deployment process of large model services and reduce the difficulty of resource configuration in the large model deployment process.

[0004] The technical solutions adopted by the present application to solve the problems are as follows:

[0005] In a first aspect, the present application provides a model deployment method based on keywords, comprising:

[0006] Obtaining a model structure keyword and a model resource keyword;

[0007] Generating a target processing model based on the model structure keyword; the target processing model includes multiple model components;

[0008] Performing resource configuration on the multiple model components in the target processing model based on the model resource keyword.

[0009] In some embodiments of the present application, after performing resource configuration on the multiple model components in the target processing model based on the model resource keyword, the method comprises:

[0010] Obtaining running state information of the multiple model components when the target processing model runs;

[0011] Performing dynamic adjustment on the resources occupied by the multiple model components based on the running state information.

[0012] In some embodiments of the present application, generating a target processing model based on a model structure keyword comprises:

[0013] Verifying the multiple model keywords in the model structure keyword and the dependency relationship between the multiple model keywords;

[0014] If the plurality of model keys and the dependency relationship among the plurality of model keys pass the check, a target processing model is generated based on the plurality of model keys and the dependency relationship among the plurality of model keys.

[0015] If the plurality of model keys and the dependency relationship among the plurality of model keys do not pass the check, first check result information is output.

[0016] An adjusted model structure key input by a user based on the first check result information is obtained, and a target processing model is generated based on the adjusted model structure key.

[0017] In some embodiments of the present application, the plurality of model keys in the model structure key and the dependency relationship among the plurality of model keys are checked, including:

[0018] Based on the plurality of model keys in the model structure key and the dependency relationship among the plurality of model keys, a corresponding model structure is determined.

[0019] If the model structure includes a preset model structure, it is determined that the plurality of model keys and the dependency relationship among the plurality of model keys do not pass the check.

[0020] If the model structure does not include a preset model structure, it is determined that the plurality of model keys and the dependency relationship among the plurality of model keys pass the check.

[0021] In some embodiments of the present application, based on the model resource key, a plurality of model components in the target processing model are resource configured, including:

[0022] The model resource key is checked.

[0023] If the model resource key passes the check, the required resource type and the required resource amount corresponding to each model component are determined from the model resource key, and the computing resources in the computing cluster are allocated to the plurality of model components in the target processing model based on the required resource type and the required resource amount.

[0024] If the model resource key does not pass the check, second check result information is output.

[0025] An adjusted model resource key input by a user based on the second check result information is obtained, and the plurality of model components in the target processing model are resource configured based on the adjusted model resource key.

[0026] In some embodiments of the present application, the model resource key includes the required resource type and the required resource amount corresponding to each model component, and the model resource key is checked, including:

[0027] matching the requirement resource type corresponding to each model component with the resource type in the computing cluster;

[0028] if the requirement resource type does not match the resource type in the computing cluster, determining that the model resource keyword check fails;

[0029] if the requirement resource type matches the resource type in the computing cluster, determining the total amount of each requirement resource type based on the requirement resource type and the requirement resource amount corresponding to each model component;

[0030] if the total amount of any requirement resource type is greater than the total amount of the requirement resource type in the computing cluster, determining that the model resource keyword check fails.

[0031] In some embodiments of the present application, the resources occupied by the plurality of model components are dynamically adjusted based on the running state information, including:

[0032] based on the running state information, determining whether the plurality of model components includes a first model component; the first model component is a model component in the plurality of model components that is running and the corresponding task is in a queuing state;

[0033] if the plurality of model components includes the first model component, determining whether the plurality of model components includes a second model component; the second model component is a model component in the plurality of model components that has not been run and occupies resources;

[0034] if the plurality of model components includes the second model component, dynamically adjusting the resources of the second model component based on the resource amount required by the task in the queuing state in the first model component.

[0035] In some embodiments of the present application, the resources of the second model component are dynamically adjusted based on the resource amount required by the task in the queuing state in the first model component, including:

[0036] if the resource amount required by the task in the queuing state in the first model component is greater than or equal to the total amount of resources occupied by the second model component, allocating the resources occupied by the second model component to the first model component;

[0037] if the resource amount required by the task in the queuing state in the first model component is less than the total amount of resources occupied by the second model component, determining the priority of the second model component based on the dependency relationship of the second model component;

[0038] determining the third model component from the second model component based on the priority and the resource amount required by the task in the queuing state in the first model component; the third model component is a model component with a higher priority in the second model component, and the total resource amount occupied by the third model component is greater than or equal to the resource amount required by the task in the queuing state in the first model component;

[0039] allocating the resource occupied by the third model component to the first model component.

[0040] In a second aspect, the present application provides a keyword-based model deployment apparatus, comprising:

[0041] an information acquisition module, configured to acquire a model structure keyword and a model resource keyword;

[0042] a model generation module, configured to generate a target processing model based on the model structure keyword; the target processing model comprises a plurality of model components;

[0043] a resource configuration module, configured to perform resource configuration on the plurality of model components in the target processing model based on the model resource keyword.

[0044] In a third aspect, the present application further provides a computer device, comprising:

[0045] one or more processors;

[0046] a memory; and

[0047] one or more application programs, wherein the one or more application programs are stored in the memory and are configured to be executed by the processor to implement the keyword-based model deployment method of any one of the first aspect.

[0048] In a fourth aspect, an embodiment of the present application provides a computer readable storage medium, having a computer program stored thereon, the computer program being loaded by a processor to execute the steps in the keyword-based model deployment method of any one of the first aspect.

[0049] The present application has the following beneficial effects: by acquiring the model structure keyword and generating the target processing model based on the model structure keyword, the computer device can automatically construct a large model service based on the model structure keyword, without the need for manual construction of the large model service by the user, thereby simplifying the deployment process of the large model service; by acquiring the model resource keyword and performing resource configuration on the plurality of model components in the target processing model based on the model resource keyword, the computer device can automatically configure resources for the model components based on the model resource keyword, thereby reducing the difficulty of resource configuration in the large model deployment process. BRIEF DESCRIPTION OF DRAWINGS

[0050] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the accompanying drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced. Obviously, the accompanying drawings in the following description only represent some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained from these accompanying drawings without creative effort.

[0051] Figure 1 is a scene schematic diagram of a keyword-based model deployment system provided by an embodiment of the present application;

[0052] Figure 2 is a flowchart of one embodiment of a keyword-based model deployment method provided by an embodiment of the present application;

[0053] Figure 3 is a structure schematic diagram of one embodiment of a target processing model provided by an embodiment of the present application;

[0054] Figure 4 is a flowchart of a specific embodiment of generating a target processing model provided by an embodiment of the present application;

[0055] Figure 5 is a structure schematic diagram of one embodiment of a model structure provided by an embodiment of the present application;

[0056] Figure 6 is a structure schematic diagram of another embodiment of a model structure provided by an embodiment of the present application;

[0057] Figure 7 is a structure schematic diagram of still another embodiment of a model structure provided by an embodiment of the present application;

[0058] Figure 8 is a structure schematic diagram of still another embodiment of a model structure provided by an embodiment of the present application;

[0059] Figure 9 is a structure schematic diagram of one embodiment of a ring model structure provided by an embodiment of the present application;

[0060] Figure 10 is a structure schematic diagram of one embodiment of a node isolated model structure provided by an embodiment of the present application;

[0061] Figure 11 is a structure schematic diagram of another embodiment of a node isolated model structure provided by an embodiment of the present application;

[0062] Figure 12 is a flowchart of a specific embodiment of performing resource configuration on a plurality of model components in a target processing model provided by an embodiment of the present application;

[0063] Figure 13is another embodiment flowchart of the keyword-based model deployment method provided by the embodiment of the present application;

[0064] Figure 14 is a running result schematic diagram of the model runtime not performing resource dynamic adjustment, provided by the embodiment of the present application;

[0065] Figure 15 is a running result schematic diagram of the model runtime performing resource dynamic adjustment, provided by the embodiment of the present application;

[0066] Figure 16 is a principle block diagram of the keyword-based model deployment device provided by the embodiment of the present application;

[0067] Figure 17 is a structure schematic diagram of one embodiment of the computer device provided by the embodiment of the present application. DETAILED DESCRIPTION

[0068] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, but not all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0069] In the description of the present application, it should be understood that the terms "first", "second", "third", etc. are used only for the purpose of description, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features limited by "first", "second", "third", etc. can be explicitly or implicitly included one or more features.

[0070] In the present application, the word "exemplary" is used to mean "serving as an example, instance, or illustration." Any implementation described as "exemplary" in the present application is not necessarily to be construed as preferred or advantageous over other implementations. The following description is presented to enable any person skilled in the art to make and use the present application. In the following description, for the purposes of explanation, numerous details are set forth. It should be appreciated, however, that the present application can be practiced in embodiments other than the embodiments implicitly illustrated or explicitly described herein. In other instances, well-known structures and processes are not described in detail in order to avoid obscuring the description of the present application. Thus, the present application is not intended to be limited to the embodiments shown, but is to be accorded the widest scope consistent with the principles and features disclosed herein.

[0071] It should be noted that the method of the embodiments of the present application is executed in a computer device, and the processing objects of each computer device exist in the form of data or information, for example, time, which is essentially time information. It can be understood that if the size, quantity, position, etc. are mentioned in the subsequent embodiments, they are corresponding data, so that the computer device can process them, and details are not described here.

[0072] The embodiments of the present application provide a keyword-based model deployment method and device, equipment and storage medium, which are described in detail below.

[0073] Please refer to Figure 1 , Figure 1 The scene schematic diagram of the keyword-based model deployment system provided by the embodiments of the present application can include a computer device 100, and the computer device 100 is integrated with a keyword-based model deployment apparatus, such as the computer device in Figure 1 .

[0074] In the embodiments of the present application, the computer device 100 is mainly used to obtain a model structure keyword and a model resource keyword; generate a target processing model based on the model structure keyword; the target processing model includes a plurality of model components; and based on the model resource keyword, the plurality of model components in the target processing model are configured with resources. The computer device can automatically construct a large model service based on the model structure keyword and automatically configure resources for the model components based on the model resource keyword, so as to simplify the deployment process of the large model service and reduce the difficulty of resource configuration in the large model deployment process.

[0075] In the embodiments of the present application, the computer device 100 can be a standalone server, or a server network or server cluster composed of servers. For example, the computer device 100 described in the embodiments of the present application includes but is not limited to a computer, a network host, a single network server, a plurality of network server sets, or a cloud server composed of a plurality of servers. The cloud server is composed of a large number of computers or network servers based on cloud computing.

[0076] It can be understood that the computer device 100 used in the embodiments of the present application can be a device that includes receiving and transmitting hardware, i.e., a device with receiving and transmitting hardware capable of performing bidirectional communication on a bidirectional communication link. Such a device can include a cellular or other communication device with a single-line display or a multi-line display or a cellular or other communication device without a multi-line display. The computer device 100 can be a desktop terminal or a mobile terminal, and the computer device 100 can also be one of a television, a mobile phone, a tablet computer, a notebook computer, etc.

[0077] Those skilled in the art will understand that Figure 1 The application environment shown is merely one application scenario of the present application solution and does not constitute a limitation on the application scenario of the present application solution. Other application environments may include more than one application scenario. Figure 1 The number of computer devices shown is more or less, for example Figure 1 Only one computer device is shown in the document. It is understood that the keyword-based model deployment system may also include one or more other services, which are not limited here.

[0078] In addition, such as Figure 1 As shown, the keyword-based model deployment system may also include a memory 200 for storing data, such as keyword information, such as model structure keywords, model resource keywords, etc., and verification result information, such as first verification result information, second verification result information, etc.

[0079] It should be noted that, Figure 1 The schematic diagram of the keyword-based model deployment system shown is merely an example. The keyword-based model deployment system and scenario described in this application are for the purpose of more clearly illustrating the technical solutions of this application and do not constitute a limitation on the technical solutions provided in this application. As those skilled in the art will know, with the evolution of keyword-based model deployment systems and the emergence of new business scenarios, the technical solutions provided in this application are also applicable to similar technical problems.

[0080] First, this application provides a keyword-based model deployment method. The execution body of the keyword-based model deployment method is a keyword-based model deployment device, which is applied to a computer device. The keyword-based model deployment method includes: obtaining model structure keywords and model resource keywords; generating a target processing model based on the model structure keywords; the target processing model includes multiple model components; and configuring resources for the multiple model components in the target processing model based on the model resource keywords.

[0081] like Figure 2 The diagram shown is a flowchart of an embodiment of the keyword-based model deployment method in this application. The keyword-based model deployment method may include steps S201 to S203, as detailed below:

[0082] S201. Obtain model structure keywords and model resource keywords.

[0083] In an embodiment of the present application, the model structure keyword is a keyword related to the model structure of the large model, and the model structure keyword includes a plurality of model keywords and a dependency relationship between the plurality of model keywords. For example, when the model structure keyword is "input: model a, model a: model b, model c, model b: model c, model d, model c: model d, model d: model e, model e: output", the model structure keyword includes seven model keywords, i.e., input, model a, model b, model c, model d, model e, and output. Model a depends on input, model b and model c depend on model a, model c and model d depend on model b, model d depends on model c, model e depends on model d, and output depends on model e. That is, model a is used to receive input information, model b and model c receive the output of model a respectively, model c and model d receive the output of model b respectively, model d receives the output of model c, model e receives the output of model d, and model e is used to output the final information.

[0084] Further, the model resource keyword is resource configuration information associated with the model components corresponding to the plurality of model keywords, and the model resource keyword includes the required resource type and the required resource amount of the model components corresponding to the plurality of model keywords. The required resource type can include any one or more of a graphics processing unit (GPU), a central processing unit (CPU), a random access memory (RAM), and a custom computing resource. For example, the model components corresponding to the plurality of model keywords include model a, model b, model c, model d, and model e, and the model resource keyword can be "model a: GPU: 2, model b: GPU: 1, model c: GPU: 0.5, model d: GPU: 0.25, model e: GPU: 0.25". That is, based on the model resource keyword, it can be determined that the required resource type of model a, model b, model c, model d, and model e is a graphics processing unit (GPU), and the required resource amount of model a, model b, model c, model d, and model e is 2, 1, 0.5, 0.25, and 0.25 respectively.

[0085] Optionally, the computer device can obtain the model structure keyword and the model resource keyword in various ways. For example, the computer device can receive the model structure keyword and the model resource keyword input by the user through a keyboard, a mouse, a touch screen, or the like. The computer device can also obtain the model structure keyword and the model resource keyword from other devices through a network, Bluetooth, or the like. The present embodiment is not limited in this regard.

[0086] S202, generate a target processing model based on the model structure keyword; the target processing model includes a plurality of model components.

[0087] In the embodiments of the present application, the target processing model is a large model service deployed based on the model structure keyword, the target processing model includes a plurality of model components, and the embodiments generate the target processing model based on the model structure keyword. The user only needs to manually input the model structure keyword, and the computer device can automatically deploy the large model service based on the model structure keyword, thereby simplifying the deployment process of the large model service.

[0088] In some embodiments, the step of generating a target processing model based on a model structure keyword specifically includes: generating a target processing model based on a plurality of model keywords in the model structure keyword and the dependency relationship between the plurality of model keywords. For example, when the model structure keyword is "input: model a, model a: model b, model c, model b: model c, model d, model c: model d, model d: model e, model e: output", a target processing model including five model components of model a, model b, model c, model d and model e can be generated based on the plurality of model keywords in the model structure keyword and the dependency relationship between the plurality of model keywords. Figure 3 As shown in the target processing model, the target processing model includes five model components of model a, model b, model c, model d and model e.

[0089] In other embodiments, referring to Figure 4 As shown in the target processing model, the target processing model includes five model components of model a, model b, model c, model d and model e.

[0090] S301, check the plurality of model keywords in the model structure keyword and the dependency relationship between the plurality of model keywords.

[0091] In some embodiments, the step of checking the plurality of model keywords in the model structure keyword and the dependency relationship between the plurality of model keywords specifically includes: determining the corresponding model structure based on the plurality of model keywords in the model structure keyword and the dependency relationship between the plurality of model keywords; if the model structure includes a preset model structure, determining that the dependency relationship between the plurality of model keywords is not checked; if the model structure does not include a preset model structure, determining that the dependency relationship between the plurality of model keywords is checked.

[0092] In the embodiments of the present application, the model structure is determined based on the plurality of model keywords and the dependency relationship between the plurality of model keywords in the model structure keyword. For example, when the model structure keyword is "model c:model d, model a:model b, model c, model b:model c, input:model a, model e:output", the model structure determined based on "model c:model d" in the model structure keyword is as shown in the following figure: Figure 5 The model structure determined based on "model a:model b, model c, model b:model c" in the model structure keyword is as shown in the following figure: Figure 6 The model structure determined based on "input:model a" in the model structure keyword is as shown in the following figure: Figure 7 The model structure determined based on "model e:output" in the model structure keyword is as shown in the following figure: Figure 8 The model structure determined based on "model e:output" in the model structure keyword is as shown in the following figure.

[0093] Further, the preset model structure is a pre-set invalid model structure, and the preset model structure can be used to measure whether the model structure is valid. Optionally, the preset model structure can include a loop model structure and a node isolated model structure. For example, referring to Figures 9 to 11 As shown in the following figure, Figure 9 The figure is an embodiment diagram of the loop model structure, in which the input and output of the plurality of model components form a loop, Figure 10 and Figure 11 The figure is an embodiment diagram of the node isolated model structure, in which the model component has only output or has only input.

[0094] In S302, if the plurality of model keywords and the dependency relationship between the plurality of model keywords pass the verification, the target processing model is generated based on the plurality of model keywords and the dependency relationship between the plurality of model keywords.

[0095] In some embodiments, if the plurality of model keywords and the dependency relationship between the plurality of model keywords pass the verification, the target processing model can be directly generated based on the plurality of model keywords and the dependency relationship between the plurality of model keywords. For example, when the model structure keyword is "input:model a, model a:model b, model c, model b:model c, model d, model c:model d, model d:model e, model e:output", the model structure determined based on the plurality of model keywords and the dependency relationship between the plurality of model keywords in the model structure keyword does not include the preset model structure, that is, the plurality of model keywords and the dependency relationship between the plurality of model keywords in the model structure keyword pass the verification, and then the target processing model can be directly generated based on the plurality of model keywords and the dependency relationship between the plurality of model keywords in the model structure keyword.

[0096] S303, if the plurality of model keys and the dependency relationship between the plurality of model keys fail to pass the verification, output first verification result information.

[0097] In the embodiments of the present application, the first verification result information is used to represent that the plurality of model keys and the dependency relationship between the plurality of model keys fail to pass the verification. Further, in order to facilitate the user to adjust the model structure key based on the first verification result information, the first verification result information can include the reason for failing to pass the verification, for example, the first verification result information is 'the model structure corresponding to'model a: model b, model b: model c, model c: model a' forms a loop, and the verification fails'.

[0098] In the embodiments of the present application, the computer device can output the first verification result information in various ways, for example, the computer device can output the first verification result information through the display screen configured by itself, the computer device can also output the first verification result information through the microphone configured by itself, and the computer device can also transmit the first verification result information to other devices through the network, Bluetooth, etc., and output the first verification result information through other devices.

[0099] S304, obtaining the adjusted model structure key input by the user based on the first verification result information, and generating a target processing model based on the adjusted model structure key.

[0100] Optionally, after the computer outputs the first verification result information, the user can adjust the model structure key based on the first verification result information, and the computer device obtains the adjusted model structure key input by the user based on the first verification result information, and generates a target processing model based on the adjusted model structure key, so as to improve the accuracy of the generated target processing model.

[0101] In some embodiments, the step of generating a target processing model based on the adjusted model structure key specifically includes: determining the adjusted model structure key as a new model structure key, and continuing to execute the step of verifying the plurality of model keys in the model structure key and the dependency relationship between the plurality of model keys until the plurality of model keys and the dependency relationship between the plurality of model keys pass the verification, and generating a target processing model based on the plurality of model keys and the dependency relationship between the plurality of model keys.

[0102] S203, based on the model resource key, performing resource configuration on the plurality of model components in the target processing model.

[0103] In this embodiment of the application, after the target processing model is generated, the resources of multiple model components in the target processing model are configured based on the model resource keywords. In this way, the computer device can automatically configure the resources of multiple model components in the target processing model based on the model resource keywords, thereby solving the problem that developers need to modify the code to manually configure model resources, which leads to high difficulty in resource configuration.

[0104] In some embodiments, the step of configuring resources for multiple model components in a target processing model based on model resource keywords specifically includes: determining the required resource type and required resource quantity for each model component from the model resource keywords, and allocating computing resources in the computing cluster to the multiple model components in the target processing model based on the required resource type and required resource quantity. For example, when the model resource keywords are "model a: GPU: 2, model b: GPU: 1, model c: GPU: 0.5, model d: GPU: 0.25, model e: GPU: 0.25", then 2 GPUs in the computing cluster are allocated to model a, 1 GPU in the computing cluster is allocated to model b, 0.5 GPUs in the computing cluster are allocated to model c, 0.25 GPUs in the computing cluster are allocated to model d, and 0.25 GPUs in the computing cluster are allocated to model e.

[0105] In other embodiments, reference is made to Figure 12 As shown, step S203 above, which configures resources for multiple model components in the target processing model based on model resource keywords, may include steps S401 to S404, as follows:

[0106] S401. Verify the keywords of the model resources.

[0107] In some embodiments, model resource keywords include the required resource type and required resource quantity corresponding to each model component. The steps for verifying the model resource keywords specifically include: matching the required resource type corresponding to each model component with the resource types in the computing cluster; if the required resource type does not match the resource types in the computing cluster, determining that the model resource keyword verification has failed; if the required resource type matches the resource types in the computing cluster, determining the total required resource quantity for each required resource type based on the required resource type and required resource quantity corresponding to each model component; if the total required resource quantity for any required resource type is greater than the total resource quantity for that required resource type in the computing cluster, determining that the model resource keyword verification has failed.

[0108] In the embodiments of the present application, the matching of the demand resource type and the resource type in the computing cluster means that the demand resource type is the resource type in the computing cluster, for example, the demand resource type is GPU, and the resource type in the computing cluster includes GPU and CPU, and thus the demand resource type is matched with the resource type in the computing cluster; otherwise, if the demand resource type is GPU, and the resource type in the computing cluster includes CPU and ARM, then the demand resource type is not matched with the resource type in the computing cluster.

[0109] Further, when determining the total demand resource amount of each demand resource type based on the demand resource type and the demand resource amount corresponding to each model component, the demand resource amounts of the same demand resource type corresponding to the plurality of model components can be added, so as to obtain the total demand resource amount of each demand resource type.

[0110] Optionally, if the demand resource type is not matched with the resource type in the computing cluster, it is determined that the model resource keyword verification fails; otherwise, if the demand resource type is matched with the resource type in the computing cluster, the total demand resource amount of each demand resource type is determined based on the demand resource type and the demand resource amount corresponding to each model component, and the verification result of the model resource keyword is determined based on the total demand resource amount of each demand resource type.

[0111] In some embodiments, the step of determining the verification result of the model resource keyword based on the total demand resource amount of each demand resource type specifically includes: comparing the total demand resource amount of each demand resource type with the total resource amount of the demand resource type in the computing cluster, and if the total demand resource amount of any demand resource type is greater than the total resource amount of the demand resource type in the computing cluster, it is determined that the model resource keyword verification fails; otherwise, if the total demand resource amount of each demand resource type is less than the total resource amount of the demand resource type in the computing cluster, it is determined that the model resource keyword verification passes. For example, the total demand resource amount of GPU is 5, and the total resource amount of GPU in the computing cluster is 4, and thus the total demand resource amount of GPU is greater than the total resource amount of GPU in the computing cluster, and the model resource keyword verification fails.

[0112] S402, if the model resource keyword verification passes, determining the demand resource type and the demand resource amount corresponding to each model component from the model resource keyword, and allocating the computing resources in the computing cluster to the plurality of model components in the target processing model based on the demand resource type and the demand resource amount.

[0113] In the embodiments of the present application, if the model resource keyword verification passes, the demand resource type and the demand resource amount corresponding to each model component are directly determined from the model resource keyword, and the computing resources in the computing cluster are allocated to the plurality of model components in the target processing model based on the demand resource type and the demand resource amount.

[0114] S403, output the second verification result information if the model resource keyword verification fails.

[0115] In the embodiments of the present application, the second verification result information is used to represent that the model resource keyword verification fails. Further, in order to facilitate the user to adjust the model resource keyword based on the second verification result information, the second verification result information can include the reason why the model resource keyword verification fails, for example, the second verification result information is "GPU resource exceeds resource total amount, verification fails", the second verification result information can also be "GPU resource demand amount is 5, GPU resource total amount is 4, GPU resource demand amount exceeds resource total amount, verification fails", and the second verification result information can also be "GPU resource type does not match, verification fails".

[0116] In the embodiments of the present application, the computer device can output the second verification result information in various ways, for example, the computer device can output the second verification result information through the display screen configured by itself, the computer device can also output the second verification result information through the microphone configured by itself, and the computer device can also transmit the second verification result information to other devices through the network, Bluetooth, etc., and output the second verification result information through other devices.

[0117] S404, obtaining the adjusted model resource keyword input by the user based on the second verification result information, and performing resource configuration on the plurality of model components in the target processing model based on the adjusted model resource keyword.

[0118] Optionally, after the computer device outputs the second verification result information, the user can adjust the model resource keyword based on the second verification result information, the computer device obtains the adjusted model resource keyword input by the user based on the second verification result information, and performs resource configuration on the plurality of model components in the target processing model based on the adjusted model resource keyword, so as to improve the accuracy of resource configuration.

[0119] In some embodiments, the step of performing resource configuration on the plurality of model components in the target processing model based on the adjusted model resource keyword specifically includes: determining the adjusted model resource keyword as a new model resource keyword, and continuing to perform the step of verifying the model resource keyword until the model resource keyword verification passes, determining the demand resource type and the demand resource amount corresponding to each model component from the model resource keyword, and allocating the computing resources in the computing cluster to the plurality of model components in the target processing model based on the demand resource type and the demand resource amount.

[0120] In some embodiments, with reference to Figure 13As shown, after the step S203 of configuring the resources of the plurality of model components in the target processing model based on the model resource keywords, the method can further include steps S204-S205, which are specifically as follows.

[0121] S204, obtaining running state information of the plurality of model components when the target processing model is running.

[0122] In the embodiments of the present application, the running state information of the plurality of model components is used to represent whether the plurality of model components are running, and the running state information of the plurality of model components can include running, and / or, has run, and / or, has not run.

[0123] S205, dynamically adjusting the resources occupied by the plurality of model components based on the running state information.

[0124] In some embodiments, the step of dynamically adjusting the resources occupied by the plurality of model components based on the running state information specifically includes: determining a fourth model component from the plurality of model components based on the running state information, the fourth model component being a model component in the plurality of model components that has run; and releasing the resources occupied by the fourth model component, so as to avoid the model component in the target processing model that has run continuing to occupy the computing resources, and improve the utilization rate of the resources.

[0125] In other embodiments, the step of dynamically adjusting the resources occupied by the plurality of model components based on the running state information specifically includes: determining whether the plurality of model components include a first model component based on the running state information; if the plurality of model components include the first model component, determining whether the plurality of model components include a second model component; and if the plurality of model components include the second model component, dynamically adjusting the resources of the second model component based on the amount of resources required by the tasks in the queuing state in the first model component.

[0126] In the embodiments of the present application, the first model component is a model component in the plurality of model components that is running and the corresponding task is in the queuing state, for example, taking the target model in the Figure 3 For example, taking the target model in the Figure 3 As shown, if the model d and the model e have not run, it is determined that the model d and the model e are the second model components.

[0127] Optionally, if the first model component is included in the plurality of model components, it indicates that the first model component is configured with insufficient resources, and the resources of the second model component are dynamically adjusted based on the resource amount required by the task in the queuing state in the first model component. The idle resources occupied by the second model component can be allocated to the first model component for task processing, so as to avoid resource waste caused by idle resources and shorten the running time of the multi-task scenario.

[0128] In some embodiments, the step of dynamically adjusting the resources of the second model component based on the resource amount required by the task in the queuing state in the first model component specifically includes: if the resource amount required by the task in the queuing state in the first model component is greater than or equal to the total amount of resources occupied by the second model component, the resources occupied by the second model component are allocated to the first model component; if the resource amount required by the task in the queuing state in the first model component is less than the total amount of resources occupied by the second model component, the priority of the second model component is determined based on the dependency relationship of the second model component; the third model component is determined from the second model component based on the priority and the resource amount required by the task in the queuing state in the first model component; and the resources occupied by the third model component are allocated to the first model component.

[0129] In the embodiments of the present application, when the priority of the second model component is determined based on the dependency relationship of the second model component, the priority of the dependent party can be set to be higher than the priority of the dependency party in the two model components having the dependency relationship, or the priority of the dependency party can be set to be higher than the priority of the dependent party in the two model components having the dependency relationship, which is not limited in the present embodiment. For example, when model b depends on model a, the priority of model a can be set to be higher than the priority of model b, or the priority of model a can be set to be lower than the priority of model b.

[0130] Further, the third model component is a model component with a high priority in the second model component, and the total amount of resources occupied by the third model component is greater than or equal to the resource amount required by the task in the queuing state in the first model component. Optionally, when the priority of the dependent party is set to be higher than the priority of the dependency party in the two model components having the dependency relationship, the third model component is a model component with a low priority in the order from low to high; when the priority of the dependency party is set to be higher than the priority of the dependent party in the two model components having the dependency relationship, the third model component is a model component with a high priority in the order from high to low. For example, when model b depends on model a, the third model component is a model component with a low priority in the order from low to high, or the third model component is a model component with a high priority in the order from high to low. Figure 3For example, if model a is a first model component, the resource required by the task in the queuing state in model a is 1 GPU, if model d and model e are both model components that are not running and occupy resources, model d and model e both occupy 1 GPU, and model e depends on model d, according to the dependency relationship, the priority of model e is higher than that of model d, and then the 1 GPU resource occupied by model e is allocated to model a to execute the task in the queuing state.

[0131] For another example, referring to FIG. 6, GPU1, GPU2 and GPU3 are allocated to model a, model b and model c respectively, model b depends on model a, and model c depends on model b, if the resources allocated to model a, model b and model c are not dynamically adjusted, GPU2 allocated to model b will be idle in the running stage of model a, and GPU3 allocated to model c will be idle in the running stages of model a and model b, which will cause waste of resources. Figure 14 Figure 15 For another example, referring to FIG. 6, GPU1, GPU2 and GPU3 are allocated to model a, model b and model c respectively, model b depends on model a, and model c depends on model b, if the resources allocated to model a, model b and model c are not dynamically adjusted, GPU2 allocated to model b will be idle in the running stage of model a, and GPU3 allocated to model c will be idle in the running stages of model a and model b, which will cause waste of resources. Figure 15

[0132] In order to better implement the model deployment method based on the keyword in the embodiments of the present application, on the basis of the model deployment method based on the keyword, the embodiments of the present application further provide a model deployment device based on the keyword, as shown in FIG. 6, the model deployment device based on the keyword 600 comprises: Figure 16

[0133] An information acquisition module 610 is configured to acquire a model structure keyword and a model resource keyword.

[0134] A model generation module 620 is configured to generate a target processing model based on the model structure keyword, and the target processing model comprises a plurality of model components.

[0135] A resource configuration module 630 is configured to perform resource configuration on the plurality of model components in the target processing model based on the model resource keyword.

[0136] ​​​In the embodiments of the present application, the model structure keyword is acquired, the target processing model is generated based on the model structure keyword, the computer device can automatically construct a large model service based on the model structure keyword, and the user does not need to manually construct the large model service, thereby simplifying the deployment process of the large model service; the model resource keyword is acquired, and the multiple model components in the target processing model are configured with resources based on the model resource keyword, the computer device can automatically configure resources for the model components based on the model resource keyword, thereby reducing the difficulty of resource configuration in the large model deployment process.

[0137] In some embodiments of the present application, the keyword-based model deployment apparatus 600 further includes:

[0138] The state acquisition module is configured to acquire running state information of the multiple model components when the target processing model runs.

[0139] The resource adjustment module is configured to dynamically adjust the resources occupied by the multiple model components based on the running state information.

[0140] In some embodiments of the present application, the model generation module 620 is specifically configured to:

[0141] verify the multiple model keywords in the model structure keyword and the dependency relationship between the multiple model keywords;

[0142] If the multiple model keywords and the dependency relationship between the multiple model keywords pass the verification, the target processing model is generated based on the multiple model keywords and the dependency relationship between the multiple model keywords.

[0143] If the multiple model keywords and the dependency relationship between the multiple model keywords do not pass the verification, the first verification result information is output.

[0144] The adjusted model structure keyword input by the user based on the first verification result information is acquired, and the target processing model is generated based on the adjusted model structure keyword.

[0145] In some embodiments of the present application, the model generation module 620 is specifically further configured to:

[0146] based on the multiple model keywords in the model structure keyword and the dependency relationship between the multiple model keywords, determine the corresponding model structure;

[0147] If the model structure includes a preset model structure, it is determined that the multiple model keywords and the dependency relationship between the multiple model keywords do not pass the verification.

[0148] If the model structure does not include a preset model structure, it is determined that the multiple model keywords and the dependency relationship between the multiple model keywords pass the verification.

[0149] In some embodiments of the present application, the resource configuration module 630 is specifically configured to:

[0150] verify the model resource keyword;

[0151] if the model resource keyword passes the verification, determine the required resource type and the required resource amount corresponding to each model component from the model resource keyword, and allocate the computing resources in the computing cluster to the plurality of model components in the target processing model based on the required resource type and the required resource amount;

[0152] if the model resource keyword fails the verification, output the second verification result information;

[0153] obtain the adjusted model resource keyword input by the user based on the second verification result information, and perform resource configuration on the plurality of model components in the target processing model based on the adjusted model resource keyword.

[0154] In some embodiments of the present application, the model resource keyword includes the required resource type and the required resource amount corresponding to each model component, and the resource configuration module 630 is specifically further configured to:

[0155] match the required resource type corresponding to each model component with the resource type in the computing cluster;

[0156] if the required resource type does not match the resource type in the computing cluster, determine that the model resource keyword fails the verification;

[0157] if the required resource type matches the resource type in the computing cluster, determine the total amount of the required resource of each required resource type based on the required resource type and the required resource amount corresponding to each model component;

[0158] if the total amount of the required resource of any required resource type is greater than the total amount of the resource of the required resource type in the computing cluster, determine that the model resource keyword fails the verification.

[0159] In some embodiments of the present application, the resource adjustment module is specifically configured to:

[0160] based on the running state information, determine whether the plurality of model components include a first model component; the first model component is a model component in the plurality of model components that is running and the corresponding task is in a queuing state;

[0161] if the plurality of model components include the first model component, determine whether the plurality of model components include a second model component; the second model component is a model component in the plurality of model components that has not been run and occupies resources;

[0162] If the second model component is included in the plurality of model components, the resource of the second model component is dynamically adjusted based on the resource amount required by the task in the queuing state in the first model component.

[0163] In some embodiments of the present application, the resource adjustment module is specifically further configured to:

[0164] If the resource amount required by the task in the queuing state in the first model component is greater than or equal to the total resource occupied by the second model component, the total resource occupied by the second model component is allocated to the first model component.

[0165] If the resource amount required by the task in the queuing state in the first model component is less than the total resource occupied by the second model component, the priority of the second model component is determined based on the dependency relationship of the second model component.

[0166] The third model component is determined from the second model component based on the priority and the resource amount required by the task in the queuing state in the first model component. The third model component is a model component with a high priority in the second model component, and the total resource occupied by the third model component is greater than or equal to the resource amount required by the task in the queuing state in the first model component.

[0167] The total resource occupied by the third model component is allocated to the first model component.

[0168] Embodiments of the present application also provide a computer device, which comprises:

[0169] One or more processors;

[0170] Memory; and

[0171] One or more application programs, wherein the one or more application programs are stored in the memory and configured to execute the steps in the keyword-based model deployment method in any one of the above-mentioned embodiments of the keyword-based model deployment method by the processor.

[0172] Embodiments of the present application also provide a computer device, as shown in Figure 17 The computer device structure involved in the embodiments of the present application is specifically shown in the structure diagram of the computer device, and specifically comprises:

[0173] The computer device can include a processor 801 with one or more processing cores, a memory 802 with one or more computer readable storage media, a power supply 803, and an input unit 804, etc. Those skilled in the art can understand that the computer device structure shown in Figure 17 does not constitute a limitation on the computer device, and can include more or fewer components than those shown, or combine certain components, or different component arrangements. Among them:

[0174] The processor 801 is the control center of the computer device, connects the various parts of the computer device through various interfaces and lines, and performs various functions and processes data of the computer device by running or executing software programs and / or modules stored in the memory 802 and calling data stored in the memory 802, thereby monitoring the computer device as a whole. Optionally, the processor 801 can include one or more processing cores; optionally, the processor 801 can integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface, and application programs, etc., and the modem processor mainly processes wireless communication. It can be understood that the above-mentioned modem processor can also not be integrated into the processor 801.

[0175] The memory 802 can be used to store software programs and modules, and the processor 801 executes various functions and data processing by running the software programs and modules stored in the memory 802. The memory 802 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application program required by a function (such as a sound playing function, an image playing function, etc.), etc.; the data storage area can store data created according to the use of the computer device, etc. In addition, the memory 802 can include a high-speed random access memory, and can also include a non-volatile memory, such as at least one magnetic disk storage device, a flash memory device, or other volatile solid-state memory device. Accordingly, the memory 802 can also include a memory controller to provide access for the processor 801 to the memory 802.

[0176] The computer device further includes a power supply 803 for powering various components, and the power supply 803 can be logically connected to the processor 801 through a power management system, so as to realize functions such as management of charging, discharging, and power consumption management through the power management system. The power supply 803 can also include one or more than one direct current or alternating current power supply, a recharging system, a power failure detection circuit, a power converter or inverter, a power state indicator, and any other components.

[0177] The computer device can also include an input unit 804, which can be used to receive input digital or character information, and generate keyboard, mouse, joystick, optical or trackball signal inputs related to user settings and function control.

[0178] Although not shown, the computer device can also include a display unit, etc., which will not be described here. In particular, in the present embodiment, the processor 801 in the computer device will load the executable file corresponding to the process of one or more than one application program into the memory 802 according to the following instructions, and run the application program stored in the memory 802 by the processor 801, thereby realizing various functions, such as:

[0179] obtain a model structure keyword and a model resource keyword;

[0180] generate a target processing model based on the model structure keyword; the target processing model comprises a plurality of model components;

[0181] perform resource configuration on the plurality of model components in the target processing model based on the model resource keyword.

[0182] Those skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructions, or by related hardware controlled by the instructions, which can be stored in a computer readable storage medium and loaded and executed by a processor.

[0183] To this end, the embodiments of the present application provide a computer readable storage medium, which can include a read only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc. A computer program is stored on the computer readable storage medium, and the computer program is loaded by a processor to execute the steps in any of the model deployment methods based on keywords provided by the embodiments of the present application. For example, the computer program loaded by the processor can execute the following steps:

[0184] obtain a model structure keyword and a model resource keyword;

[0185] generate a target processing model based on the model structure keyword; the target processing model comprises a plurality of model components;

[0186] perform resource configuration on the plurality of model components in the target processing model based on the model resource keyword.

[0187] In the above embodiments, the description of each embodiment has its own focus, and the parts not described in detail in a certain embodiment can be referred to the detailed description of other embodiments above, which will not be repeated here.

[0188] In a specific implementation, each of the above units or structures can be implemented as an independent entity, or can be combined as the same or several entities, and the specific implementation of each of the above units or structures can be referred to the method embodiments above, which will not be repeated here.

[0189] The specific implementation of each of the above operations can be referred to the embodiments above, which will not be repeated here.

[0190] The above describes in detail the keyword-based model deployment method, device, equipment and storage medium provided by the embodiments of the present application. The principles and implementation manners of the present application are described by applying specific examples. The above description of the embodiments is only used to help understand the method of the present application and its core idea. Meanwhile, for those skilled in the art, according to the idea of the present application, the specific implementation manners and application ranges will be changed. In conclusion, the content of the specification should not be understood as a limitation of the present application.

Claims

1. A keyword-based model deployment method, characterized in that, include: Obtain model structure keywords and model resource keywords; Based on the aforementioned model structure keywords, a target processing model is generated; The target processing model is a large model service deployed based on model structure keywords, and the target processing model includes multiple model components; Based on the model resource keywords, resource configuration is performed on multiple model components in the target processing model; After configuring resources for multiple model components in the target processing model based on the model resource keywords, the process includes: When the target processing model is running, the running status information of multiple model components is obtained; the running status information of multiple model components is used to characterize whether multiple model components are running, and the running status information of multiple model components includes running, and / or, already running, and / or, not running; The resources occupied by multiple model components are dynamically adjusted based on the aforementioned operational status information; The dynamic adjustment of resources occupied by multiple model components based on the running status information includes: Based on the running status information, it is determined whether the first model component is included among the multiple model components; the first model component is the model component that is currently running and whose corresponding task is in a queue state among the multiple model components. If the first model component is included in the plurality of model components, determine whether the second model component is included in the plurality of model components; the second model component is the model component that is not yet running and occupies resources among the plurality of model components; If the second model component is included in multiple model components, the resources of the second model component are dynamically adjusted based on the amount of resources required by the tasks in the queue of the first model component; The dynamic adjustment of the resources of the second model component based on the resource requirements of the tasks in the queue in the first model component includes: If the amount of resources required by the tasks in the queue in the first model component is greater than or equal to the total amount of resources occupied by the second model component, the resources occupied by the second model component will be allocated to the first model component. If the amount of resources required by the tasks in the queue in the first model component is less than the total amount of resources occupied by the second model component, the priority of the second model component is determined based on the dependency relationship of the second model component; when determining the priority of the second model component based on the dependency relationship of the second model component, the priority of the dependent component in the two model components with a dependency relationship is set to be higher than the priority of the dependent component, or the priority of the dependent component in the two model components with a dependency relationship is set to be higher than the priority of the dependent component. A third model component is determined from the second model components based on the priority and the resource requirements of the queued tasks in the first model component. The third model component is the model component with the highest priority among the second model components, and the total resources occupied by the third model component are greater than or equal to the resource requirements of the queued tasks in the first model component. When the priority of the dependent component is higher than the priority of the dependent component in two model components with a dependency relationship, the third model component is the model component with the highest priority in ascending order. When the priority of the dependent component is higher than the priority of the dependent component in two model components with a dependency relationship, the third model component is the model component with the highest priority in descending order. The resources allocated to the dependent component in the multiple model components have a certain idle time during the running phase of the dependent component. The resources occupied by the third model component are allocated to the first model component.

2. The keyword-based model deployment method according to claim 1, characterized in that, The step of generating the target processing model based on the model structure keywords includes: The multiple model keywords in the model structure keywords and the dependencies between the multiple model keywords are verified. If the multiple model keywords and the dependencies between them pass the verification, a target processing model is generated based on the multiple model keywords and their dependencies. If the validation of multiple model keywords and the dependency relationships between multiple model keywords fails, the first validation result information is output. Obtain the adjusted model structure keywords input by the user based on the first verification result information, and generate the target processing model based on the adjusted model structure keywords.

3. The keyword-based model deployment method according to claim 2, characterized in that, The verification of multiple model keywords in the model structure keywords and the dependencies between the multiple model keywords includes: Based on multiple model keywords in the model structure keywords and the dependencies between the multiple model keywords, the corresponding model structure is determined; If the model structure contains a preset model structure, the verification of multiple model keywords and the dependency relationships between the multiple model keywords fails. If the model structure does not contain the preset model structure, the verification of multiple model keywords and the dependencies between multiple model keywords passes.

4. The keyword-based model deployment method according to claim 1, characterized in that, The step of configuring resources for multiple model components in the target processing model based on the model resource keywords includes: The keywords of the model resources are validated; If the model resource keyword verification passes, the required resource type and required resource quantity for each model component are determined from the model resource keyword, and the computing resources in the computing cluster are allocated to multiple model components in the target processing model based on the required resource type and the required resource quantity. If the model resource keyword verification fails, output the second verification result information; Obtain the adjusted model resource keywords input by the user based on the second verification result information, and configure resources for multiple model components in the target processing model based on the adjusted model resource keywords.

5. The keyword-based model deployment method according to claim 4, characterized in that, The model resource keywords include the required resource type and required resource quantity for each model component; The validation of the model resource keywords includes: Match the required resource type of each model component with the resource type in the computing cluster; If the required resource type does not match the resource type in the computing cluster, the model resource keyword verification is determined to have failed. If the required resource type matches the resource type in the computing cluster, the total required resource amount for each required resource type is determined based on the required resource type and required resource amount corresponding to each model component. If the total amount of resources required for any of the resource types exceeds the total amount of resources for that resource type in the computing cluster, the model resource keyword verification is determined to have failed.

6. A keyword-based model deployment device, characterized in that, include: The information acquisition module is used to acquire model structure keywords and model resource keywords; The model generation module is used to generate a target processing model based on the model structure keywords; The target processing model is a large model service deployed based on model structure keywords, and the target processing model includes multiple model components; The resource configuration module is used to configure resources for multiple model components in the target processing model based on the model resource keywords. The keyword-based model deployment device further includes: The status acquisition module is used to acquire the running status information of multiple model components when the target processing model is running; the running status information of multiple model components is used to characterize whether multiple model components are running, and the running status information of multiple model components includes running, and / or, already running, and / or, not running; The resource adjustment module is used to dynamically adjust the resources occupied by multiple model components based on the running status information; The resource adjustment module is also specifically used for: Based on the running status information, it is determined whether the first model component is included among the multiple model components; the first model component is the model component that is currently running and whose corresponding task is in a queue state among the multiple model components. If the first model component is included in the plurality of model components, determine whether the second model component is included in the plurality of model components; the second model component is the model component that is not yet running and occupies resources among the plurality of model components; If the second model component is included in multiple model components, the resources of the second model component are dynamically adjusted based on the amount of resources required by the tasks in the queue of the first model component; The resource adjustment module is also specifically used for: If the amount of resources required by the tasks in the queue in the first model component is greater than or equal to the total amount of resources occupied by the second model component, the resources occupied by the second model component will be allocated to the first model component. If the amount of resources required by the tasks in the queue in the first model component is less than the total amount of resources occupied by the second model component, the priority of the second model component is determined based on the dependency relationship of the second model component; when determining the priority of the second model component based on the dependency relationship of the second model component, the priority of the dependent component in the two model components with a dependency relationship is set to be higher than the priority of the dependent component, or the priority of the dependent component in the two model components with a dependency relationship is set to be higher than the priority of the dependent component. A third model component is determined from the second model components based on the priority and the resource requirements of the queued tasks in the first model component. The third model component is the model component with the highest priority among the second model components, and the total resources occupied by the third model component are greater than or equal to the resource requirements of the queued tasks in the first model component. When the priority of the dependent component is higher than the priority of the dependent component in two model components with a dependency relationship, the third model component is the model component with the highest priority in ascending order. When the priority of the dependent component is higher than the priority of the dependent component in two model components with a dependency relationship, the third model component is the model component with the highest priority in descending order. The resources allocated to the dependent component in the multiple model components have a certain idle time during the running phase of the dependent component. The resources occupied by the third model component are allocated to the first model component.

7. A computer device, characterized in that, The computer device includes: One or more processors; Memory; and One or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the processor to implement the keyword-based model deployment method of any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, It stores a computer program, which is loaded by a processor to perform the steps in the keyword-based model deployment method according to any one of claims 1 to 5.

Images