Model delivery method, model delivery platform, equipment and storage medium

Abstract

The invention provides a model delivery method, a model delivery platform, equipment and a storage medium. The model delivery method comprises the following steps: acquiring model meta-information; analyzing the model meta-information to obtain training data; training the machine learning model based on the training data to obtain a trained target machine learning model; analyzing the model meta-information to obtain a model environment corresponding to the machine learning model; deploying a target machine learning model based on the model environment; and under the condition that deployment of the target machine learning model is completed, starting the target machine learning model. In the embodiment of the invention, in a model training process, training data is acquired through model meta-information; in the model deployment process, the model environment is obtained through the model meta-information, so that model training and model deployment can be realized without cross-platform, and the model delivery efficiency is improved.

Description

Technical Field

[0001] This application relates to the field of machine learning technology, and particularly to a model delivery method, a model delivery platform, a device, and a storage medium. Background Art

[0002] The general process of AI model service from requirement submission, to model development and training, and then to delivery and online launch is as follows: The data processor processes the training data, the algorithm engineer develops a machine learning model and trains the machine learning model using the training data. After the machine learning model is trained, the operation and maintenance engineer deploys the environment and deploys the trained machine learning model, and then performs online inference through the machine learning model to achieve model delivery.

[0003] Based on this, the development and training process of the machine learning model is completed by the algorithm engineer, and the deployment process of the machine learning model is completed by the operation and maintenance engineer. Since the algorithm engineer and the operation and maintenance engineer work through their respective development platforms, that is, model delivery requires operations across multiple platforms, which results in low model delivery efficiency. Summary of the Invention

[0004] The main purpose of this application is to provide a model delivery method, a model delivery platform, a device, and a storage medium, aiming to solve the technical problem that model delivery requires operations across multiple platforms, which results in low model delivery efficiency.

[0005] To achieve the above purpose, this application provides a model delivery method, including: Obtain model meta-information; Parse the model meta-information to obtain training data; Train a machine learning model based on the training data to obtain a trained target machine learning model; Parse the model meta-information to obtain the model environment corresponding to the machine learning model; Deploy the target machine learning model based on the model environment; Enable the target machine learning model when the deployment of the target machine learning model is completed.

[0006] Optionally, before obtaining the model meta-information, the method includes: Import a data set; the data set includes multiple candidate data; Perform data annotation on each candidate data; Delete the abnormal data in the multiple candidate data to obtain training data; the data annotation corresponding to the abnormal data indicates that the candidate data is abnormal; Store the training data in the model meta-information.

[0007] Optionally, before obtaining the model meta-information, the method includes: Obtain the model development code; Compile the model development code to generate a machine learning model; Obtain the model environment code; Debug the training environment corresponding to the machine learning model based on the model environment code; Store the model development code and the model environment code into the model meta-information; Wherein, the model development code is used to compile the machine learning model, and the model environment code is used to compile the model environment corresponding to the machine learning model.

[0008] Optionally, deploying the target machine learning model based on the model environment includes: Receive a computing power adjustment instruction and a capacity adjustment instruction; In response to the computing power adjustment instruction, reallocate the computing power resources corresponding to the target machine learning model in the target cluster; the target cluster is the cluster where the target machine learning model is deployed; In response to the capacity adjustment instruction, reallocate the storage resources corresponding to the target machine learning model in the target cluster; Deploy the target machine learning model with reallocated computing power resources and storage resources in the model environment.

[0009] Optionally, the method further includes: Receive an account login instruction; the account login instruction carries an account name and an account password; In response to the account login instruction, verify the account name and the account password; When the verification is passed and a read instruction is received, read the model meta-information associated with the account name; Wherein, the model meta-information includes training data, model development code, and model environment code.

[0010] In addition, to achieve the above object, the present application further provides a model delivery platform, including a model training module, a model deployment module, and a model inference module. The model training module is communicatively connected to the model deployment module, and the model deployment module is communicatively connected to the model inference module: The model training module is used to obtain model meta-information; Parse the model meta-information to obtain training data; Train a machine learning model based on the training data to obtain a trained target machine learning model; A model deployment module, configured to parse the model meta-information to obtain a model environment corresponding to a machine learning model; Deploy the target machine learning model based on the model environment; A model inference module, configured to enable the target machine learning model when the deployment of the target machine learning model is completed.

[0011] Optionally, the model delivery platform further includes: A data processing module, configured to import a data set; the data set includes a plurality of candidate data; Perform data annotation on each candidate data; Delete abnormal data from the plurality of candidate data to obtain training data; the data annotation corresponding to the abnormal data indicates that the candidate data is abnormal; Store the training data into the model meta-information.

[0012] Optionally, the model delivery platform further includes: A model development module, configured to obtain model development code; Compile the model development code to generate a machine learning model; Obtain model environment code; Debug the training environment corresponding to the machine learning model based on the model environment code; Store the model development code and the model environment code into the model meta-information; Wherein, the model development code is used to compile the machine learning model, and the model environment code is used to compile the model environment corresponding to the machine learning model.

[0013] To solve the above technical problems, an embodiment of the present application further provides a computer device, which adopts the following technical solutions: The computer device includes a memory and a processor. A computer program is stored in the memory. When the processor executes the computer program, the steps of any one of the model delivery methods proposed in the embodiments of the present application are implemented.

[0014] To solve the above technical problems, an embodiment of the present application further provides a computer-readable storage medium, which adopts the following technical solutions: A computer program is stored on the computer-readable storage medium. When the computer program is executed by a processor, the steps of any one of the model delivery methods proposed in the embodiments of the present application are implemented.

[0015] Compared with the prior art, the embodiments of the present application mainly have the following beneficial effects: The present application provides a model delivery method, a model delivery platform, a device, and a storage medium. The model delivery method includes: obtaining model meta-information; parsing the model meta-information to obtain training data; training a machine learning model based on the training data to obtain a trained target machine learning model; parsing the model meta-information to obtain a model environment corresponding to the machine learning model; deploying the target machine learning model based on the model environment; and enabling the target machine learning model when the deployment of the target machine learning model is completed. In the embodiments of the present application, during the model training process, training data is obtained through the model meta-information; during the model deployment process, the model environment is obtained through the model meta-information, so that model training and model deployment can be realized without crossing platforms, improving the efficiency of model delivery. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the solutions in the present application, the following will briefly introduce the drawings required for the description of the embodiments of the present application. Obviously, the following-described drawings are some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0017] Figure 1 is an exemplary system architecture diagram to which the present application can be applied; Figure 2 is a flowchart of the model delivery method provided by the embodiments of the present application; Figure 3 is a schematic diagram of an application scenario of the model delivery method provided by the embodiments of the present application; Figure 4 is a schematic diagram of the structure of the model delivery platform provided by the embodiments of the present application; Figure 5 is a schematic diagram of the structure of the model delivery platform provided by the embodiments of the present application; Figure 6 is a basic structural block diagram of a computer device provided by the embodiments of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] The model delivery method provided by the embodiments of the present application is applied to a model delivery platform. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present application belongs; the terms used in the specification of the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above drawings are intended to cover non-exclusive inclusion. The terms "first", "second", etc. in the specification and claims of the present application or the above drawings are used to distinguish different objects and are not used to describe a specific order.

[0019] As used herein, the term "embodiment" means that the specific features, structures, or characteristics described in connection with the embodiment may be included in at least one embodiment of the present application. The phrase appears in various places in the specification and does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art will explicitly and implicitly understand that the embodiments described herein may be combined with other embodiments.

[0020] In order to enable those skilled in the art to better understand the solution of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0021] As Figure 1 shown, the system architecture 100 may include terminal devices 101, 102, 103, a network 104, and a server 105. The network 104 is used to provide a medium for communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links, or fiber optic cables, etc.

[0022] Users may use the terminal devices 101, 102, 103 to interact with the server 105 through the network 104 to receive or send messages, etc. Various communication client applications may be installed on the terminal devices 101, 102, 103, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social online platform software, etc.

[0023] The terminal devices 101, 102, 103 may be various electronic devices with a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, e-book readers, MP3 players (Moving Picture Experts Group Audio Layer III), MP4 (Moving Picture Experts Group Audio Layer IV) players, laptop computers, and desktop computers, etc.

[0024] The server 105 may be a server providing various services, such as a background server supporting the pages displayed on the terminal devices 101, 102, 103.

[0025] It should be noted that the model delivery method provided by the embodiments of the present application is generally executed by the server / terminal device. Correspondingly, the model delivery platform is generally set in the server / terminal device.

[0026] It should be understood that Figure 1 the numbers of the terminal devices, networks, and servers in Figure 1 are merely illustrative. According to the implementation requirements, there can be any number of terminal devices, networks, and servers.

[0027] Please refer to Figure 2 , which shows a flowchart of an embodiment of the model delivery method proposed according to this application. The embodiments of this application can acquire and process relevant data based on artificial intelligence technology.

[0028] The model delivery method provided by the embodiments of this application is applied to a model delivery platform, which includes a data processing module, a model development module, a model training module, a model deployment module, and a model inference module. Optionally, the above model delivery platform is a machine learning platform. The model delivery method includes the following steps: S210, acquire model meta-information.

[0029] S220, parse the model meta-information to obtain training data.

[0030] It should be noted that the model delivery platform pre-stores model meta-information. The above model meta-information can be stored in the database of the model delivery platform, or the model delivery platform can also acquire model meta-information through blockchain technology.

[0031] In this step, after acquiring the model meta-information, parse the model meta-information to obtain training data.

[0032] S230, train a machine learning model based on the training data to obtain a trained target machine learning model.

[0033] In this step, after obtaining the training data, the training data can be used to train a machine learning model, and then a trained machine learning model is obtained. The above trained machine learning model is called a target machine learning model.

[0034] Optionally, the above model meta-information further includes model development code. In this step, the model meta-information can be parsed to obtain the model development code, and the model development code is compiled to obtain a machine learning model.

[0035] S240, parse the model meta-information to obtain the model environment corresponding to the machine learning model.

[0036] S250, deploy the target machine learning model based on the model environment.

[0037] In this step, the model meta-information can be parsed to obtain the model environment corresponding to the machine learning model, and the target machine learning model is deployed in the model environment.

[0038] Optionally, the above model meta-information further includes a model environment code. In this step, the model environment code can be parsed from the model meta-information, and the model environment can be obtained by compiling the model environment code.

[0039] S260, when the target machine learning model is deployed successfully, enable the target machine learning model.

[0040] In this step, when the target machine learning model is deployed successfully, the target machine learning model can be put online and enabled. Optionally, the target machine learning model can be used to perform inference on the received service information.

[0041] In the embodiments of the present application, model meta-information is obtained; the model meta-information is parsed to obtain training data; the machine learning model is trained based on the training data to obtain a trained target machine learning model; the model environment corresponding to the machine learning model is parsed from the model meta-information; the target machine learning model is deployed based on the model environment; when the target machine learning model is deployed successfully, the target machine learning model is enabled. In the embodiments of the present application, during the model training process, the training data is obtained through the model meta-information; during the model deployment process, the model environment is obtained through the model meta-information, so that model training and model deployment can be achieved without cross-platform, improving the efficiency of model delivery.

[0042] Optionally, before obtaining the model meta-information, the method includes: Import a data set; the data set includes a plurality of candidate data; Perform data annotation on each candidate data; Delete the abnormal data in the plurality of candidate data to obtain training data; the data annotation corresponding to the abnormal data indicates that the candidate data is abnormal; Store the training data in the model meta-information.

[0043] As described above, the model delivery platform includes a data processing module.

[0044] In this embodiment, the data processing module imports a data set in response to a data processing request. Among them, the above data set can be data stored locally on the machine learning platform, data obtained by the machine learning platform through the Internet, or data obtained by the machine learning platform through communication with other devices.

[0045] The data processing module performs data annotation on each candidate data in the data set in response to a data annotation instruction.

[0046] In an alternative implementation, the candidate data is manually labeled. In this implementation, the data processing module sends multiple candidate data included in the data set to the front-end page, and the data processor sends a data labeling instruction to the data processing module based on the candidate data to achieve manual data labeling of the multiple candidate data.

[0047] In another alternative implementation, the candidate data is automatically labeled. In this implementation, the data processing module performs structured processing on the candidate data to obtain structured data, and performs data labeling on the multiple candidate data based on preset labeling rules.

[0048] After data labeling for each candidate data, based on the labeling information corresponding to each candidate data, abnormal data in the multiple candidate data is deleted to obtain training data, and the training data is stored in the model meta-information.

[0049] In this embodiment, data is imported through the data processing module, and the data is labeled to obtain training data, and then the training data is stored in the model meta-information. Since the model meta-information can be transmitted across platforms between the model development module, the model training module, the model deployment module, and the model inference module, during the model training process, there is no need to copy the training data generated by the data processing module to the model training module, thereby improving the efficiency of model delivery.

[0050] Optionally, before obtaining the model meta-information, the method includes: Obtain the model development code; Compile the model development code to generate a machine learning model; Obtain the model environment code; Debug the training environment corresponding to the machine learning model based on the model environment code; Store the model development code and the model environment code in the model meta-information; Wherein, the model development code is used to compile the machine learning model, and the model environment code is used to compile the model environment corresponding to the machine learning model.

[0051] As described above, the model delivery platform includes a model development module and a model deployment module.

[0052] In this embodiment, the model development module obtains the model development code input by the algorithm engineer and compiles the model development code to generate a machine learning model. The model deployment module obtains the model environment code input by the algorithm engineer and debugs the training environment corresponding to the machine learning model based on the model environment code. Further, the model development code and the model environment code are stored in the model meta-information.

[0053] In this embodiment, the model development code and the model environment code are stored in the model meta-information. Since the model meta-information can be transmitted across platforms among the model development module, the model training module, the model deployment module, and the model inference module, during the model deployment process, there is no need to repeatedly set up the model environment and repeatedly compile the machine learning model, thereby improving the efficiency of model delivery.

[0054] To facilitate understanding of the technical solutions described in all the above embodiments, please refer to Figure 3 , Figure 3 The "Data Preparation" part shown can be understood as the data processing module in the embodiments of the present application, the "Model Development" part can be understood as the model development module in the embodiments of the present application, the "Model Training" part can be understood as the model training module in the embodiments of the present application, the "Model Deployment" part can be understood as the model deployment module in the embodiments of the present application, and the "Inference" part can be understood as the model inference module in the embodiments of the present application.

[0055] As Figure 3 shown, the data processing module responds to the instructions sent by the data processor, and the model development module, the model training module, the model deployment module, and the model inference module respond to the instructions sent by the algorithm engineer. The model meta-information can be transmitted among the model processing module, the model development module, the model training module, the model deployment module, and the model inference module.

[0056] In Figure 3 the application scenario shown, both model training and model deployment are completed by the algorithm engineer, and model training and model deployment can be achieved without cross-platform, thereby improving the efficiency of model delivery.

[0057] Optionally, the deploying the target machine learning model based on the model environment includes: Receiving a computing power adjustment instruction and a capacity adjustment instruction; In response to the computing power adjustment instruction, reallocating the computing power resources corresponding to the target machine learning model in the target cluster; the target cluster is the cluster where the target machine learning model is deployed; In response to the capacity adjustment instruction, reallocating the storage resources corresponding to the target machine learning model in the target cluster; Deploying the target machine learning model with reallocated computing power resources and storage resources in the model environment.

[0058] During the model deployment process, if a computing power adjustment instruction is received, based on this computing power adjustment instruction, reallocate the computing power resources corresponding to the target machine learning model. Specifically, a mapping relationship can be established between the target machine learning model and the computing nodes in the target cluster, and the target machine learning model is deployed through the computing nodes with the mapping relationship.

[0059] During the process of model deployment, if a capacity adjustment instruction is received, based on this capacity adjustment instruction, reallocate the storage resources corresponding to the target machine learning model. Specifically, the target machine learning model can be scaled up or down.

[0060] In this embodiment, during the process of model deployment, by means of a computing power adjustment instruction and a capacity adjustment instruction, adjust the computing power resources and storage resources of the target machine learning model to achieve one-click deployment of the model and improve the convenience of model deployment.

[0061] Optionally, the method further includes: Receive an account login instruction; the account login instruction carries an account name and an account password; In response to the account login instruction, verify the account name and the account password; When the verification is passed and a read instruction is received, read the model meta-information associated with the account name; Wherein, the model meta-information includes training data, model development code, and model environment code.

[0062] In this embodiment, the model delivery platform can also set personal accounts, where the personal accounts are associated with the model meta-information, and different accounts are associated with different model meta-information.

[0063] Optionally, the method is to receive an account login instruction, parse the account login instruction to obtain the account name and the account password, and verify the account name and the account password.

[0064] An optional verification method is to query the account name in a preset mapping table to obtain the password. If the password is consistent with the account password, it is determined that the verification is passed. Among them, the mapping table stores the mapping relationship between the account name and the password.

[0065] When the verification is passed, if a read instruction indicating to obtain model meta-information is received, obtain the model meta-information associated with the account name.

[0066] In this embodiment, by verifying the account, the model meta-information corresponding to the account is obtained, so as to realize the customization of the personal working directory of each account and enrich the personalized model delivery process.

[0067] Please refer to Figure 4 , a model delivery platform 400 provided by an embodiment of the present application. The model delivery platform 400 includes a model training module 410, a model deployment module 420, and a model inference module 430. The model training module 410 is communicatively connected to the model deployment module 420, and the model deployment module 420 is communicatively connected to the model inference module 430: A model training module 410, configured to obtain model meta-information; Parse the model meta-information to obtain training data; Train a machine learning model based on the training data to obtain a trained target machine learning model; A model deployment module 420, configured to parse the model meta-information to obtain a model environment corresponding to the machine learning model; Deploy the target machine learning model based on the model environment; A model inference module 430, configured to enable the target machine learning model when the deployment of the target machine learning model is completed.

[0068] In the embodiment of the present application, the above-mentioned model training module 410, model deployment module 420, and model inference module 430 can be operated by an algorithm engineer. That is, the algorithm engineer can implement the processes of model training, model deployment, and model inference within a single platform. In this way, model training and model deployment can be achieved without crossing platforms, improving the efficiency of model delivery.

[0069] Please refer to Figure 5 , a model delivery platform 400 provided by an embodiment of the present application. The model delivery platform 400 further includes: A data processing module 440, configured to import a data set; the data set includes a plurality of candidate data; Perform data annotation on each candidate data; Delete abnormal data in the plurality of candidate data to obtain training data; the data annotation corresponding to the abnormal data indicates that the candidate data is abnormal; Store the training data into the model meta-information.

[0070] Please refer to Figure 5 , a model delivery platform 400 provided by an embodiment of the present application. The model delivery platform 400 further includes: A model development module 450, configured to obtain model development code; Compile the model development code to generate a machine learning model; Obtain model environment code; Debug the training environment corresponding to the machine learning model based on the model environment code; Store the model development code and the model environment code into the model meta-information; wherein, the model development code is used to compile the machine learning model, and the model environment code is used to compile the model environment corresponding to the machine learning model.

[0071] Optionally, the model deployment module 420 is specifically configured to: Receive a computing power adjustment instruction and a capacity adjustment instruction; In response to the computing power adjustment instruction, reallocate the computing power resources corresponding to the target machine learning model in the target cluster; the target cluster is the cluster where the target machine learning model is deployed; In response to the capacity adjustment instruction, reallocate the storage resources corresponding to the target machine learning model in the target cluster; Deploy the target machine learning model with reallocated computing power resources and storage resources in the model environment.

[0072] Optionally, the model delivery platform 400 is further configured to: Receive an account login instruction; the account login instruction carries an account name and an account password; In response to the account login instruction, verify the account name and the account password; When the verification is passed and a read instruction is received, read the model meta-information associated with the account name; Wherein, the model meta-information includes training data, model development code, and model environment code.

[0073] To solve the above technical problems, an embodiment of the present application further provides a computer device. For details, please refer to Figure 6 , Figure 6 This is the basic structural block diagram of the computer device in this embodiment. The computer device 6 includes a memory 61, a processor 62, and a network interface 63 that are communicatively connected to each other through a system bus. It should be noted that only the computer device 6 with a memory 61, a processor 62, and a network interface 63 is shown in the figure, but it should be understood that it is not required to implement all the shown components, and more or fewer components can be alternatively implemented. Among them, those skilled in the art of the present technology can understand that a computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to a microprocessor, an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA), a digital signal processor (DSP), an embedded device, etc.

[0074] The computer device can be a desktop computer, a notebook, a palm computer, a cloud server, and other computing devices. The computer device can perform human-computer interaction with the user through a keyboard, a mouse, a remote control, a touchpad, a voice control device, and other means.

[0075] The memory 61 at least includes one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (such as SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 61 may be an internal storage unit of the computer device 6, such as the hard disk or memory of the computer device 6. In other embodiments, the memory 61 may also be an external storage device of the computer device 6, such as a plug-in hard disk, Smart Media Card (SMC), Secure Digital (SD) card, FlashCard, etc. equipped on the computer device 6. Of course, the memory 61 may also include both the internal storage unit and the external storage device of the computer device 6. In this embodiment, the memory 61 is generally used to store the operating system and various application software installed on the computer device 6, such as the program code of the model delivery method. In addition, the memory 61 may also be used to temporarily store various data that have been output or will be output.

[0076] In some embodiments, the processor 62 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chips. The processor 62 is generally used to control the overall operation of the computer device 6. In this embodiment, the processor 62 is used to run the program code stored in the memory 61 or process data, such as running the program code of the model delivery method.

[0077] The network interface 63 may include a wireless network interface or a wired network interface, and the network interface 63 is generally used to establish a communication connection between the computer device 6 and other electronic devices.

[0078] The present application also provides another implementation manner, that is, to provide a computer-readable storage medium storing the model delivery program, and the model delivery program can be executed by at least one processor, so that the at least one processor executes the steps of the model delivery method as described above.

[0079] Through the description of the above embodiments, those skilled in the art can clearly understand that the above-described example methods can be implemented by means of software plus a necessary general hardware online platform. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation. Based on such an understanding, the technical solution of the present application, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions for causing a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in various embodiments of the present application.

[0080] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multi-processor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and so on. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in a distributed computing environment where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0081] Obviously, the above-described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. The accompanying drawings show preferred embodiments of the present application, but do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosed content of the present application more thorough and comprehensive. Although the present application has been described in detail with reference to the foregoing embodiments, for those skilled in the art, they can still modify the technical solutions described in the foregoing specific embodiments, or perform equivalent replacements on some of the technical features. Any equivalent structure directly or indirectly using the content of the specification and drawings of the present application in other related technical fields is similarly within the scope of the patent protection of the present application.

Claims

1. A model delivery method, characterized in that: include: Get model meta information; Parsing the model meta information to obtain training data; Training the machine learning model based on the training data to obtain a trained target machine learning model; Parsing the model meta-information to obtain a model environment corresponding to the machine learning model; Deploy the target machine learning model based on the model environment; When the target machine learning model is deployed, enable the target machine learning model.

2. The method according to claim 1, characterized in that Before obtaining the model meta-information, the method includes: Importing a data set, wherein the data set includes a plurality of candidate data; Label each candidate data; Deleting abnormal data from the plurality of candidate data to obtain training data; the data annotation corresponding to the abnormal data indicates that the candidate data is abnormal; The training data is stored in the model meta information.

3. The method according to claim 1, characterized in that Before obtaining the model meta-information, the method includes: Get the model development code; Compiling the model development code to generate a machine learning model; Get the model environment code; Debugging the training environment corresponding to the machine learning model based on the model environment code; storing the model development code and the model environment code into model meta information; Among them, the model development code is used to compile the machine learning model, and the model environment code is used to compile the model environment corresponding to the machine learning model.

4. The method according to claim 1, characterized in that: The deploying the target machine learning model based on the model environment includes: Receive computing power adjustment instructions and capacity adjustment instructions; In response to the computing power adjustment instruction, reallocate computing power resources corresponding to the target machine learning model in a target cluster; the target cluster is a cluster where the target machine learning model is deployed; In response to the capacity adjustment instruction, reallocate storage resources corresponding to the target machine learning model in the target cluster; Deploy a target machine learning model in the model environment and reallocate computing resources and storage resources.

5. The method according to any one of claims 1 to 4, characterized in that The method further comprises: Receive an account login instruction; the account login instruction carries an account name and an account password; In response to the account login instruction, verifying the account name and the account password; If the verification is successful and a read instruction is received, read the model meta information associated with the account name; The model meta-information includes training data, model development code and model environment code.

6. A model delivery platform, characterized in that: It includes a model training module, a model deployment module and a model reasoning module, wherein the model training module is communicatively connected with the model deployment module, and the model deployment module is communicatively connected with the model reasoning module: Model training module, used to obtain model meta-information; Parsing the model meta information to obtain training data; Training the machine learning model based on the training data to obtain a trained target machine learning model; A model deployment module, used to parse the model meta-information to obtain a model environment corresponding to the machine learning model; Deploy the target machine learning model based on the model environment; The model inference module is used to enable the target machine learning model when the target machine learning model is deployed.

7. The model delivery platform according to claim 6, characterized in that: The model delivery platform also includes: A data processing module, used for importing a data set; the data set includes a plurality of candidate data; Label each candidate data; Deleting abnormal data from the plurality of candidate data to obtain training data; the data annotation corresponding to the abnormal data indicates that the candidate data is abnormal; The training data is stored in the model meta information.

8. The model delivery platform according to claim 6, characterized in that: The model delivery platform also includes: Model development module, used to obtain model development code; Compiling the model development code to generate a machine learning model; Get the model environment code; Debugging the training environment corresponding to the machine learning model based on the model environment code; storing the model development code and the model environment code into model meta information; Among them, the model development code is used to compile the machine learning model, and the model environment code is used to compile the model environment corresponding to the machine learning model.

9. A computer device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the model delivery method according to any one of claims 1 to 5 when executing the computer program.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the model delivery method according to any one of claims 1 to 5 are implemented.

Images