WO2023097952A1 - 预训练模型发布方法及装置、电子设备、存储介质和计算机程序产品 - Google Patents

预训练模型发布方法及装置、电子设备、存储介质和计算机程序产品 Download PDF

Info

Publication number
WO2023097952A1
WO2023097952A1 PCT/CN2022/088585 CN2022088585W WO2023097952A1 WO 2023097952 A1 WO2023097952 A1 WO 2023097952A1 CN 2022088585 W CN2022088585 W CN 2022088585W WO 2023097952 A1 WO2023097952 A1 WO 2023097952A1
Authority
WO
WIPO (PCT)
Prior art keywords
model
training
information
trained
update
Prior art date
Application number
PCT/CN2022/088585
Other languages
English (en)
French (fr)
Inventor
吴珂馨
杨冠姝
曾天
张弛
杨阳
Original Assignee
上海商汤智能科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 上海商汤智能科技有限公司 filed Critical 上海商汤智能科技有限公司
Publication of WO2023097952A1 publication Critical patent/WO2023097952A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/44—Arrangements for executing specific programs
    • G06F9/451—Execution arrangements for user interfaces
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches

Definitions

  • the present disclosure relates to but not limited to the field of computer technology, and in particular relates to a pre-training model publishing method and device, electronic equipment, storage media and computer program products.
  • the pre-trained model can be obtained through pre-training.
  • the model will be trained on a large amount of common corpus, and the general language knowledge will be learned to obtain the pre-trained model.
  • the pre-trained model is targeted for migration training to obtain a model suitable for the target task.
  • the pre-training models obtained by different users do not have uniform specifications and standards, it is not convenient for other users to use the pre-training models, resulting in low usability and convenience of the pre-training models.
  • Embodiments of the present disclosure provide a method and device for releasing a pre-training model, electronic equipment, a storage medium, and a computer program product.
  • An embodiment of the present disclosure provides a method for publishing a pre-trained model, including:
  • An embodiment of the present disclosure provides a device for publishing a pre-trained model, including:
  • the first upload part is configured to receive the pre-training model uploaded by the target object, and display the model information input interface;
  • the first information receiving part is configured to receive the model information of the pre-training model input by the target object on the input interface
  • the first verification part is configured to verify the model information according to a preset standard
  • the first release part is configured to release the pre-trained model when the model information meets the standard.
  • An embodiment of the present disclosure provides an electronic device, including: a processor; a memory configured to store instructions executable by the processor; wherein the processor is configured to invoke the instructions stored in the memory to execute the above method.
  • An embodiment of the present disclosure provides a computer-readable storage medium, on which computer program instructions are stored, and when the computer program instructions are executed by a processor, the foregoing method is implemented.
  • An embodiment of the present disclosure provides a computer program product, where the computer program product includes a computer program or an instruction, and when the computer program or instruction is run on an electronic device, the electronic device is made to execute the foregoing method.
  • the model information input interface is displayed by receiving the pre-training model uploaded by the target object; receiving the model information of the pre-training model input by the target object on the input interface; The model information is verified; when the model information conforms to the standard, the pre-training model is released.
  • a standardized pre-training model release process is provided, and the model information is verified through preset standards, so that the released pre-training model meets the preset standards, reducing the adaptation cost caused by inconsistent model information.
  • Fig. 1 shows a flowchart of a pre-trained model release method according to an embodiment of the present disclosure.
  • Fig. 2 shows a schematic diagram of a model information input interface according to an embodiment of the present disclosure.
  • Fig. 3 shows a schematic diagram of the relationship between states of a pre-training model according to an embodiment of the present disclosure.
  • Fig. 4 shows a block diagram of an apparatus for publishing a pre-trained model according to an embodiment of the present disclosure.
  • Fig. 5 shows a block diagram of an electronic device according to an embodiment of the present disclosure.
  • Fig. 6 shows a block diagram of an electronic device according to an embodiment of the present disclosure.
  • the pre-training model release method provided by the embodiments of the present disclosure performs unified and standardized centralized management on the pre-training model, and the user can download and call the released pre-training model for testing and tuning.
  • the model information input interface is displayed by receiving the pre-training model uploaded by the target object; receiving the model information of the pre-training model input by the target object on the input interface; The model information is verified; when the model information conforms to the standard, the pre-training model is released.
  • a standardized pre-training model release process is provided, and the model information is verified through preset standards, so that the released pre-training model meets the preset standards, reducing the adaptation cost caused by inconsistent model information.
  • the execution subject of the pre-training model release method may be a pre-training model management platform or an electronic device, and the platform or electronic device may run on an electronic device such as a terminal device or a server.
  • the pre-training model publishing method can also be executed by electronic devices such as terminal equipment or servers, and the terminal equipment can be user equipment (User Equipment, UE), mobile equipment, user terminal, terminal, cellular phone, cordless phone, personal digital assistant (Personal Digital Assistant) Assistant, PDA), handheld device, computing device, vehicle-mounted device, wearable device, etc., the method can be realized by calling the computer-readable instructions stored in the memory by the processor.
  • the execution subject of the pre-training model release method may be the pre-training model management platform. way to introduce. It can be understood that the execution subject of the method is the pre-training model management platform is only an exemplary description, and should not be understood as a limitation of the method.
  • Fig. 1 shows a flowchart of a pre-training model release method provided by an embodiment of the present disclosure.
  • the pre-training model release method includes:
  • step S11 the pre-training model uploaded by the target object is received, and the model information input interface is displayed;
  • the target audience may be the user who invented the pre-trained model.
  • the target object can upload the pre-training model on the pre-training model management platform, and the pre-training model uploaded by the target object can be stored in the memory for subsequent release.
  • the pre-trained model uploaded by the target object can be a model that realizes any function, for example, it can be a model for image recognition, a model for speech recognition, or a model for text recognition. There are no specific restrictions on the public function of the pre-trained model uploaded by the target object.
  • the model information input interface is used for the target object to input the model information of the pre-training model.
  • the model information includes information describing the pre-training model. For details, please refer to the possible implementation manners provided by the embodiments of the present disclosure.
  • the upload model page appears for the target object to upload the pre-trained model and enter the model information.
  • step S12 receiving the model information of the pre-training model input by the target object on the input interface
  • the target object After displaying the model information input interface to the target object, the target object can input the model information of the pre-training model in the interface, and submit it to the pre-training model management platform, and the management platform can receive the information of the pre-training model input by the target object model information.
  • step S13 the model information is verified according to preset standards
  • the preset standard can be a preset standard for standardizing model information, and the preset standard can include that the model information contains all required information, the warehouse address stored in the model training code in the model information can be accessed, and the model information contains The configuration file and structure file of the model can be visualized normally, or the training performance indicators of the pre-trained model are normal.
  • step S14 when the model information conforms to the standard, the pre-trained model is released.
  • the target object can fill in the version number of the released pre-training model. After confirming the release, the model will be made public. After the pre-training model is released, other users can download and test the pre-training model or tuning.
  • publishing the pre-training model may be publishing the pre-training model to a specified user-accessible platform, and publishing the pre-training model may also be setting the pre-training model to a specified user-accessible state , the specified users here can be some users or any users.
  • the embodiment of the present disclosure by receiving the pre-training model and the model information of the pre-training model uploaded by the target object, the model information is verified according to the preset standard, and the pre-training model is released when the model information meets the standard . Therefore, the embodiment of the present disclosure provides a standardized pre-training model publishing process, and checks the model information through preset standards, so that the released pre-training model meets the preset standards, reducing the risk caused by inconsistency of model information. Adaptation costs, improving the usability and convenience of released pre-trained models.
  • the model information includes at least one of the following: model name, model training task category, warehouse address where the model training code is stored, the business item to which the model belongs, the size of the storage space occupied by the model, and the configuration of the model files, model structure files, and model training performance metrics.
  • FIG. 2 is a schematic diagram of a model information input interface provided by the present disclosure.
  • model name (name) is used to name the pre-trained model, and the model name is a required item.
  • the training task category (task) of the model is used to represent the category of the training task of the model.
  • the category can be a detection task category, and the detection task category is the category to which the model that performs the target detection task belongs, for example, In a model for detecting faces in images, or in a model for detecting cats in images.
  • the category may also be a classification task category, and the classification task category is a category to which a model performing a classification task belongs, for example, a model used to classify images.
  • the warehouse address (git repo) where the model training code is stored is used to indicate the storage location of the model training program in the network.
  • the model training program is a program used to train the pre-trained model.
  • the program uses training samples to train the pre-trained model. train.
  • the business item (tag) to which the model belongs can be a certain business item in the enterprise, that is, which business item in the enterprise the pre-trained model belongs to.
  • the size of the storage space occupied by the model (model structure), which is used to characterize the storage space of the memory occupied by the model.
  • the size of the storage space occupied by the model can be a specific value of the storage space, for example: 5G; the storage space occupied by the model
  • the size of the space may also be a word used to roughly express the size of the storage space, for example, it may be small, medium, large, and so on.
  • the configuration file (config) of the model contains the configuration information of the model parameters.
  • the model structure file contains parameters used to characterize the model structure, such as weight parameters of nodes in the model.
  • the training performance index (metric) of the model is used to characterize the performance of the model on a certain training set, for example, it can be the accuracy of the model on each sample set and so on.
  • model information may also include other additional information.
  • the three fields of the warehouse address where the model training code is stored, the configuration file of the model, and the structure file of the model can uniquely define a pre-trained model, so when changing the model information, these three fields are not allowed to be modified .
  • users of the pre-training model can better adapt the model, especially for the company, reducing the inconsistency of model information within the company , resulting in adaptation costs.
  • the preset standard includes at least one of the following: the model information contains all required information; the warehouse address stored in the model training code in the model information can be accessed; the model The configuration file and model structure file in the information can be visualized normally; the training performance indicators of the pre-trained model are normal.
  • the required information in the model information can be specified.
  • the required information can be the model name, the training task category of the model, the warehouse address where the model training code is stored, and the business item to which the model belongs , the configuration file of the model, and the structure file of the model.
  • access detection can also be performed on the warehouse address stored in the model training code in the model information to determine whether it can be accessed; it can also be checked whether the configuration file in the model information and the structure file of the model can Normal visualization, so that you can view and edit configuration files and structure files correctly; you can also verify whether the training performance indicators of the pre-trained model are normal.
  • access detection can also be performed on the warehouse address stored in the model training code in the model information to determine whether it can be accessed; it can also be checked whether the configuration file in the model information and the structure file of the model can Normal visualization, so that you can view and edit configuration files and structure files correctly; you can also verify whether the training performance indicators of the pre-trained model are normal.
  • the accuracy rate in the performance index is greater than 100%, it can be determined that the training performance index of the model is abnormal and the verification fails.
  • the model information is verified through the preset standard, so that the released pre-training model meets the preset standard, which reduces the adaptation cost caused by the inconsistency of the model information, and improves the release of pre-training.
  • Model availability and convenience By viewing the model information, users can determine the performance of the model to decide whether to use the model; and the warehouse address stored in the model training code can be accessed normally, so that users who use the model can call the model normally for subsequent model tuning .
  • the receiving the pre-training model uploaded by the target object includes: receiving the pre-training model uploaded by the target object to the background management platform; the publishing the pre-training model includes: publishing the pre-training model The model is published to the default platform.
  • the preset platform may be any public platform accessible to users.
  • the pre-training model can be divided into two pieces of data.
  • the pre-training model is stored in the background management platform.
  • the pre-training model in the background management platform can be seen by the uploading user himself or the background The administrator user in the management platform can see it; and when the pre-training model is in the published state, a copy of the pre-training model will be copied to the public platform for storage to improve data security.
  • the pre-training model in the public platform It is available for users to download, test and tune.
  • the method further includes: receiving an update model uploaded by the target object for updating the published pre-training model, displaying a model information input interface to the target object; receiving the target object input update the model information of the model; verify the model information of the updated model according to preset standards, and obtain a verification result; according to the verification result, determine that the model information of the updated model meets the standard, and The updated model is published.
  • the target object can be supported to update the published pre-training model
  • the update process is similar to the process of uploading a new pre-training model
  • the model information input interface can also be displayed to the target object.
  • the object enters the model information of the updated model, and the model information of the updated model can refer to the relevant description of the previous model information; after the target object submits the model information, it can receive the model information of the updated model input by the target object, and then update the model according to the preset standard
  • the model information of the model is verified, and the verification process of obtaining the verification result can refer to the relevant description of the model information verification above; to publish.
  • the updated model may be released to the public platform, and the user can download, test, and tune the updated model.
  • the update model can also be divided into two pieces of data.
  • the update model is stored in the background management platform, which can be seen by the uploading user himself, and can also be seen by the background management platform. Visible to administrator users; while the pre-training model on the public platform is a previously public pre-training model, when the updated model is in the released state, the updated model will be copied to the public platform for storage to improve data security , the updated model on the public platform is available for users to download, test, and tune.
  • the pre-training model release method further includes: comparing the update model with the pre-training model, determining update information between the update model and the pre-training model, and generating Instructions for updating the pre-trained model version.
  • the update information of this version (such as a release notice) can be displayed in each area of the interface, that is, the version is updated relative to the previous version At the same time, it can also display the model information of this version.
  • the update information between the update model and the pre-training model can be determined by comparing the update model and the pre-training model, that is, the difference between the update model and the pre-training model can be determined. After the update information is determined, the update information can be displayed as a version update description for the pre-trained model.
  • update information between the update model and the pre-training model is determined by comparing the update model with the pre-training model; a pre-training model version update description is generated according to the update information .
  • the state of the pre-training model includes at least one of the following: an unpublished state, a published state, an updated state, and an offline state;
  • the operations supported by the pre-trained model in the unreleased state include : Publish, edit, view details, delete;
  • the operations supported by the pre-trained model in the published state include: edit, view details, and unpublish;
  • the operations supported by the pre-trained model in the updated state include: publish, edit, View details and cancel publishing;
  • the operations supported by the offline pre-trained model include: delete.
  • Figure 3 is a schematic diagram of the relationship between the states of a pre-training model provided by the present disclosure.
  • the target object after the target object uploads the pre-training model, the pre-training model is in the unpublished state 301 , at this time, the target object can perform operations such as publishing, editing, viewing details, and deleting the pre-training model; after the pre-training model is released, the pre-training model is in the released state 302, and the target object can edit and view the details of the pre-training model , cancel publishing and other operations; after the target object unpublishes the pre-training model in the released state 302, the pre-training model will enter the offline state 303, and for the pre-training model in the offline state 303, the target object can The training model performs a delete operation; after the target object updates and publishes the pre-trained model in the published state 302, the updated model enters the updated state 304, and the model in the updated state 304 can support the target object to perform release, edit, and view details
  • the embodiment of the present disclosure also provides a pre-training model release device, electronic equipment, computer-readable storage medium and computer program product, all of which can be used to implement any pre-training model release method provided by the embodiment of the present disclosure, and the corresponding technical solutions and description and see corresponding notes in the methods section.
  • FIG. 4 shows a block diagram of a pre-training model release device provided by an embodiment of the present disclosure. As shown in FIG. 4, the device 20 includes:
  • the first uploading part 21 is configured to receive the pre-training model uploaded by the target object, and display the model information input interface;
  • the first information receiving part 22 is configured to receive the model information of the pre-training model input by the target object on the input interface
  • the first verification part 23 is configured to verify the model information according to a preset standard
  • the first release part 24 is configured to release the pre-trained model when the model information meets the standard.
  • the first uploading part is also configured to receive the pre-training model uploaded by the target object to the background management platform; the first publishing part is also configured to release the pre-training model to the preset platform .
  • the model information includes at least one of the following:
  • Model name Model name, model training task category, warehouse address where model training code is stored, business project to which the model belongs, size of storage space occupied by the model, model configuration file, model structure file, and model training performance indicators.
  • the preset criteria include at least one of the following:
  • the warehouse address stored in the model training code in the model information can be accessed
  • the configuration file in the model information and the structure file of the model can be visualized normally;
  • the training performance indicators of the pre-trained model are normal.
  • the device also includes:
  • the second uploading part is configured to receive an update model uploaded by the target object and is configured to update the published pre-training model, and display a model information input interface of the update model;
  • the second information receiving part is configured to receive the model information of the updated model input by the target object
  • the second verification part is configured to verify the model information of the updated model according to a preset standard, and obtain a verification result
  • the second release part is configured to release the update model when it is determined that the model information of the update model conforms to the standard according to the verification result.
  • the device also includes:
  • the comparison part is configured to compare the updated model with the pre-trained model, and determine update information between the updated model and the pre-trained model;
  • the update description generation part is configured to generate a pre-training model version update description according to the update information.
  • the state of the pre-trained model includes at least one of the following:
  • the operations supported by the pre-trained model in the unpublished state include: publishing, editing, viewing details, and deleting;
  • the operations supported by the pre-trained model in the published state include: editing, viewing details, canceling publishing;
  • the operations supported by the pre-trained model in the updated state include: release, edit, view details, cancel release;
  • the operations supported by the offline pre-trained model include: delete.
  • the functions or parts included in the apparatus provided in the embodiments of the present disclosure may be configured to execute the methods described in the above method embodiments, and for specific implementation, refer to the descriptions of the above method embodiments.
  • a "part" may be a part of a circuit, a part of a processor, a part of a program or software, etc., of course it may also be a unit, a module or a non-modular one.
  • Embodiments of the present disclosure also provide a computer-readable storage medium, on which computer program instructions are stored, and the above-mentioned method is implemented when the computer program instructions are executed by a processor.
  • Computer readable storage media may be volatile or nonvolatile computer readable storage media.
  • An embodiment of the present disclosure also proposes an electronic device, including: a processor; a memory configured to store instructions executable by the processor; wherein the processor is configured to invoke the instructions stored in the memory to execute the above method.
  • An embodiment of the present disclosure also provides a computer program product, including computer-readable codes, or a non-volatile computer-readable storage medium carrying computer-readable codes, when the computer-readable codes are stored in a processor of an electronic device When running in the electronic device, the processor in the electronic device executes the above method.
  • an electronic device may be provided as a terminal, a server, or other forms of devices.
  • FIG. 5 shows a block diagram of an electronic device 800 provided by an embodiment of the present disclosure.
  • the electronic device 800 may be a user equipment (User Equipment, UE), a mobile device, a user terminal, a terminal, a cellular phone, a cordless phone, a personal digital assistant (PDA), a handheld device, a computing device, a vehicle-mounted device, Terminal devices such as wearable devices.
  • UE User Equipment
  • PDA personal digital assistant
  • Terminal devices such as wearable devices.
  • electronic device 800 may include one or more of the following components: processing component 802, memory 804, power supply component 806, multimedia component 808, audio component 810, input/output (Input/Output, I/O) interface 812, sensor component 814, and communication component 816.
  • the processing component 802 generally controls the overall operations of the electronic device 800, such as those associated with display, telephone calls, data communications, camera operations, and recording operations.
  • the processing component 802 may include one or more processors 820 to execute instructions to complete all or part of the steps of the above method. Additionally, processing component 802 may include one or more components that facilitate interaction between processing component 802 and other components. For example, processing component 802 may include a multimedia portion to facilitate interaction between multimedia component 808 and processing component 802 .
  • the memory 804 is configured to store various types of data to support operations at the electronic device 800 . Examples of such data include instructions of any application or method configured to operate on the electronic device 800, contact data, phonebook data, messages, pictures, videos, and the like.
  • the memory 804 can be realized by any type of volatile or non-volatile storage device or their combination, such as Static Random-Access Memory (Static Random-Access Memory, SRAM), Electrically Erasable Programmable Read-Only Memory (Electrically Erasable Programmable Read Only Memory, EEPROM), Erasable Programmable Read Only Memory (Erasable Programmable Read Only Memory, EPROM), Programmable Read Only Memory (Programmable Read Only Memory, PROM), Read Only Memory (Read Only Memory, ROM ), magnetic memory, flash memory, magnetic or optical disk.
  • Static Random-Access Memory Static Random-Access Memory
  • SRAM Static Random-Access Memory
  • EEPROM Electrically Erasable Programmable Read-Only Memory
  • Erasable Programmable Read Only Memory Eras
  • the power supply component 806 provides power to various components of the electronic device 800 .
  • Power components 806 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for electronic device 800 .
  • the multimedia component 808 includes a screen providing an output interface between the electronic device 800 and the user.
  • the screen may include a liquid crystal display (Liquid Crystal Display, LCD) and a touch panel (Touch Panel, TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from a user.
  • the touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensor may not only sense a boundary of a touch or swipe action, but also detect duration and pressure associated with the touch or swipe action.
  • the multimedia component 808 includes a front camera and/or a rear camera. When the electronic device 800 is in an operation mode, such as a shooting mode or a video mode, the front camera and/or the rear camera can receive external multimedia data. Each front camera and rear camera can be a fixed optical lens system or have focal length and optical zoom capability.
  • the audio component 810 is configured to output and/or input audio signals.
  • the audio component 810 includes a microphone (Microphone, MIC), and when the electronic device 800 is in an operation mode, such as a calling mode, a recording mode and a voice recognition mode, the microphone is configured to receive an external audio signal. Received audio signals may be further stored in memory 804 or sent via communication component 816 .
  • the audio component 810 also includes a speaker configured to output audio signals.
  • the I/O interface 812 provides an interface between the processing component 802 and the peripheral interface part, which may be a keyboard, a click wheel, buttons, and the like. These buttons may include, but are not limited to: a home button, volume buttons, start button, and lock button.
  • Sensor assembly 814 includes one or more sensors configured to provide various aspects of status assessment for electronic device 800 .
  • the sensor component 814 can detect the open or closed state of the electronic device 800, the relative positioning of components, such as the display and the keypad of the electronic device 800, the sensor component 814 can also detect the electronic device 800 or the electronic device 800 Changes in the position of components, presence or absence of user contact with electronic device 800 , electronic device 800 orientation or acceleration or deceleration and temperature changes in electronic device 800 .
  • Sensor assembly 814 may include a proximity sensor configured to detect the presence of nearby objects in the absence of any physical contact.
  • the sensor assembly 814 may also include an optical sensor, such as a Complementary Metal Oxide Semiconductor Sensor (CMOS) or Charge-Coupled Device (CCD) image sensor, configured for use in imaging applications.
  • CMOS Complementary Metal Oxide Semiconductor Sensor
  • CCD Charge-Coupled Device
  • the sensor component 814 may also include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor or a temperature sensor.
  • the communication component 816 is configured to facilitate wired or wireless communication between the electronic device 800 and other devices.
  • the electronic device 800 can access wireless networks based on communication standards, such as wireless networks (Wireless Fidelity, Wi-Fi), second generation mobile communication technologies (2nd Generation, 2G), third generation mobile communication technologies (3rd Generation, 3G) , the fourth generation of mobile communication technology (4th Generation, 4G), the long-term evolution of general mobile communication technology (Long Term Evolution, LTE), the fifth generation of mobile communication technology (5th Generation, 5G) or their combination.
  • the communication component 816 receives broadcast signals or broadcast related information from an external broadcast management system via a broadcast channel.
  • the communication component 816 also includes a Near Field Communication (NFC) portion to facilitate short-range communication.
  • NFC Near Field Communication
  • the NFC part can be based on radio frequency identification (Radio Frequency Identification, RFID) technology, infrared data association (Infrared-Data-Association, IrDA) technology, ultra-wideband (Ultra Wide Band, UWB) technology, Bluetooth (Bluetooth, BT) technology and other technologies to achieve.
  • RFID Radio Frequency Identification
  • IrDA infrared data association
  • UWB ultra-wideband
  • Bluetooth Bluetooth
  • the electronic device 800 may be implemented by one or more application-specific integrated circuits (Application-Specific Integrated Circuit, ASIC), digital signal processors (Digital System Processor, DSP), digital signal processing equipment (Digital Signal Processing Devices, DSPD), Programmable Logic Device (Programmable Logic Device, PLD), Field Programmable Gate Array (Field Programmable Gate Array, FPGA), controller, microcontroller, microprocessor or other electronic components, configured to execute the above method.
  • ASIC Application-Specific Integrated Circuit
  • DSP Digital System Processor
  • DSPD digital signal processing equipment
  • PLD Programmable Logic Device
  • Field Programmable Gate Array Field Programmable Gate Array
  • FPGA Field Programmable Gate Array
  • a non-volatile computer-readable storage medium such as the memory 804 including computer program instructions, which can be executed by the processor 820 of the electronic device 800 to implement the above method.
  • the embodiments of the present disclosure relate to the field of augmented reality.
  • the target object may involve faces, limbs, gestures, actions, etc. related to the human body, or markers and markers related to objects, or sand tables, display areas or display items related to venues or places.
  • Vision-related algorithms can involve visual positioning, Simultaneous Localization and Mapping (SLAM), 3D reconstruction, image registration, background segmentation, key point extraction and tracking of objects, pose or depth detection of objects, etc.
  • SLAM Simultaneous Localization and Mapping
  • Specific applications can not only involve interactive scenes such as guided tours, navigation, explanations, reconstructions, virtual effect overlays and display related to real scenes or objects, but also special effects processing related to people, such as makeup beautification, body beautification, special effect display, virtual Interactive scenarios such as model display.
  • the relevant features, states and attributes of the target object can be detected or identified through the convolutional neural network.
  • the above-mentioned convolutional neural network is a network model obtained by performing model training based on a deep learning framework.
  • FIG. 6 shows a block diagram of an electronic device 1900 according to an embodiment of the present disclosure.
  • electronic device 1900 may be provided as a server.
  • electronic device 1900 includes processing component 1922 , which further includes one or more processors, and a memory resource represented by memory 1932 for storing instructions executable by processing component 1922 , such as application programs.
  • An application program stored in memory 1932 may include one or more portions each corresponding to a set of instructions.
  • the processing component 1922 is configured to execute instructions to perform the above method.
  • Electronic device 1900 may also include a power component 1926 configured to perform power management of electronic device 1900 , a wired or wireless network interface 1950 configured to connect electronic device 1900 to a network, and an input-output (I/O) interface 1958 .
  • the electronic device 1900 can operate based on the operating system stored in the memory 1932, such as the Microsoft server operating system (Windows Server TM), the graphical user interface-based operating system (Mac OS XTM) introduced by Apple Inc., and the multi-user and multi-process computer operating system ( Unix TM), a free and open source Unix-like operating system (Linux TM), an open source Unix-like operating system (FreeBSD TM), or similar.
  • Microsoft server operating system Windows Server TM
  • Mac OS XTM graphical user interface-based operating system
  • Unix TM multi-user and multi-process computer operating system
  • Linux TM free and open source Unix-like operating system
  • FreeBSD TM open source Unix-like operating system
  • non-volatile computer-readable storage medium such as the memory 1932 including computer program instructions, which can be executed by the processing component 1922 of the electronic device 1900 to complete the above method.
  • Embodiments of the present disclosure may be systems, methods and/or computer program products.
  • a computer program product may include a computer-readable storage medium carrying computer-readable program instructions configured to cause a processor to implement various aspects of embodiments of the present disclosure.
  • a computer readable storage medium may be a tangible device that can retain and store instructions for use by an instruction execution device.
  • a computer readable storage medium may be, for example, but is not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing.
  • Computer-readable storage media include: portable computer diskettes, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), or flash memory), static random access memory (SRAM), portable compact disk read-only memory (Compact Disc Read-Only Memory, CD-ROM), digital versatile disk (Digital Versatile Disc, DVD), memory stick, floppy disk, mechanical Encoding devices, such as punched cards or raised structures in grooves having instructions stored thereon, and any suitable combination of the foregoing.
  • RAM random access memory
  • ROM read-only memory
  • EPROM erasable programmable read-only memory
  • flash memory static random access memory
  • SRAM static random access memory
  • portable compact disk read-only memory Compact Disc Read-Only Memory
  • CD-ROM Compact Disc Read-Only Memory
  • DVD digital versatile disk
  • memory stick floppy disk
  • mechanical Encoding devices such as punched cards or raised structures in grooves having instructions stored thereon, and any suitable combination of the foregoing.
  • computer-readable storage media are not to be construed as transient signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., pulses of light through fiber optic cables), or transmitted electrical signals.
  • Computer-readable program instructions described herein may be downloaded from a computer-readable storage medium to a respective computing/processing device, or downloaded to an external computer or external storage device over a network, such as the Internet, a local area network, a wide area network, and/or a wireless network.
  • the network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and/or edge servers.
  • a network adapter card or a network interface in each computing/processing device receives computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in a computer-readable storage medium in each computing/processing device .
  • Computer program instructions configured to perform operations of embodiments of the present disclosure may be assembly instructions, Instruction Set Architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or in the form of a or any combination of programming languages, including object-oriented programming languages—such as Smalltalk, C++, etc., and conventional procedural programming languages—such as “C” or similar programming languages language.
  • Computer readable program instructions may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer, or entirely on the remote computer or server implement.
  • the remote computer can be connected to the user's computer through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or it can be connected to an external computer such as use an Internet service provider to connect via the Internet).
  • LAN Local Area Network
  • WAN Wide Area Network
  • an Internet service provider to connect via the Internet.
  • computer-readable program instructions to personalize and customize electronic circuits, such as programmable logic circuits, field programmable gate arrays (FPGA) or programmable logic arrays (Programmable Array Logic, PLA).
  • the electronic circuitry can execute computer readable program instructions to implement various aspects of the present disclosure.
  • These computer-readable program instructions may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine such that when executed by the processor of the computer or other programmable data processing apparatus , producing an apparatus for realizing the functions/actions specified in one or more blocks in the flowchart and/or block diagram.
  • These computer-readable program instructions can also be stored in a computer-readable storage medium, and these instructions cause computers, programmable data processing devices and/or other devices to work in a specific way, so that the computer-readable medium storing instructions includes An article of manufacture comprising instructions for implementing various aspects of the functions/acts specified in one or more blocks in flowcharts and/or block diagrams.
  • each block in a flowchart or block diagram may represent a portion, a program segment, or a portion of instructions that includes one or more Executable instructions.
  • the functions noted in the block may occur out of the order noted in the figures. For example, two blocks in succession may, in fact, be executed substantially concurrently, or they may sometimes be executed in the reverse order, depending upon the functionality involved.
  • each block of the block diagrams and/or flowchart illustrations, and combinations of blocks in the block diagrams and/or flowchart illustrations can be implemented by a dedicated hardware-based system that performs the specified function or action , or may be implemented by a combination of dedicated hardware and computer instructions.
  • the computer program product can be specifically realized by means of hardware, software or a combination thereof.
  • the computer program product is embodied as a computer storage medium, and in another optional embodiment, the computer program product is embodied as a software product, such as a software development kit (Software Development Kit, SDK) etc. wait.
  • a software development kit Software Development Kit, SDK
  • Embodiments of the present disclosure provide a pre-training model release method and device, electronic equipment, storage media, and computer program products, which receive a pre-training model uploaded by a target object and display a model information input interface; receive the target object in the input The model information of the pre-training model input through the interface; the model information is verified according to the preset standard; and the pre-training model is released when the model information meets the standard.
  • a standardized pre-training model release process is provided, and the model information is verified through preset standards, so that the released pre-training model meets the preset standards, reducing the adaptation cost caused by inconsistent model information.
  • Improved usability and convenience of published pretrained models improved usability and convenience of published pretrained models.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • Software Systems (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Computation (AREA)
  • Evolutionary Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Human Computer Interaction (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

本公开实施例提供一种预训练模型发布方法及装置、电子设备、存储介质和计算机程序产品，所述方法应用于电子设备，所述方法包括：接收目标对象上传的预训练模型，显示模型信息输入界面；接收所述目标对象在所述输入界面输入的所述预训练模型的模型信息；根据预设的标准对所述模型信息进行核验；在所述模型信息符合所述标准的情况下，对所述预训练模型进行发布。

Description

预训练模型发布方法及装置、电子设备、存储介质和计算机程序产品
相关申请的交叉引用
本公开实施例基于申请号为202111440634.5、申请日为2021年11月30日、申请名称为“预训练模型发布方法及装置、电子设备和存储介质”的中国专利申请提出，并要求该中国专利申请的优先权，该中国专利申请的全部内容在此引入本公开作为参考。
技术领域
本公开涉及但不限于计算机技术领域，尤其涉及一种预训练模型发布方法及装置、电子设备、存储介质和计算机程序产品。
背景技术
随着信息技术的不断发展，人工智能技术越来越普及，模型训练作为人工智能技术的重要组成也越来越受到重视。在模型训练的过程中，可以通过预训练得到预训练模型，在预训练过程中，会先在大量通用的语料上训练模型，学习到通用的语言知识，得到预训练模型。然后，针对不同的任务，再针对性地对预训练模型进行迁移训练，得到适用于目标任务的模型。
由于不同用户预训练得到的预训练模型没有统一的规范和标准，不便于其它用户对预训练模型的使用，导致预训练模型的可用性和便利性不高。
发明内容
本公开实施例提出了一种预训练模型发布方法及装置、电子设备、存储介质和计算机程序产品。
本公开实施例提供一种预训练模型发布方法，包括：
接收目标对象上传的预训练模型，显示模型信息输入界面；
接收所述目标对象在所述输入界面输入的所述预训练模型的模型信息；
根据预设的标准对所述模型信息进行核验；
在所述模型信息符合所述标准的情况下，对所述预训练模型进行发布。
本公开实施例提供一种预训练模型发布装置，包括：
第一上传部分，配置为接收目标对象上传的预训练模型，显示模型信息输入界面；
第一信息接收部分，配置为接收所述目标对象在所述输入界面输入的所述预训练模型的模型信息；
第一核验部分，配置为根据预设的标准对所述模型信息进行核验；
第一发布部分，配置为在所述模型信息符合所述标准的情况下，对所述预训练模型进行发布。
本公开实施例提供一种电子设备，包括：处理器；配置为存储处理器可执行指令的存储器；其中，所述处理器配置为调用所述存储器存储的指令，以执行上述方法。
本公开实施例提供一种计算机可读存储介质，其上存储有计算机程序指令，所述计算机程序指令被处理器执行时实现上述方法。
本公开实施例提供一种计算机程序产品，所述计算机程序产品包括计算机程序或指令，在所述计算机程序或指令在电子设备上运行的情况下，使得所述电子设备执行上述方法。
在本公开实施例中，通过接收目标对象上传的预训练模型，显示模型信息输入界面；接收所述目标对象在所述输入界面输入的所述预训练模型的模型信息；根据预设的标准对所述模型信息进行核验；在所述模型信息符合所述标准的情况下，对所述预训练模型进行发布。由此，提供了规范化的预训练模型发布流程，并通过预设的标准对模型信息进行核验，使得发布的预训练模型符合预设的标准，减少了因模型信息的不一致导致的适配成本，提高了发布的预训练模型的可用性和便利性。
应当理解的是，以上的一般描述和后文的细节描述仅是示例性和解释性的，而非限制本公开。根据下面参考附图对示例性实施例的详细说明，本公开的其它特征及方面将变得清楚。
附图说明
为了更清楚地说明本公开实施例的技术方案，下面将对本公开实施例中所需要使用的附图进行说明。
此处的附图被并入说明书中并构成本说明书的一部分，这些附图示出了符合本公开的实施例，并与说明书一起用于说明本公开的技术方案。
图1示出根据本公开实施例的预训练模型发布方法的流程图。
图2示出根据本公开实施例的一种模型信息输入界面的示意图。
图3示出根据本公开实施例的一种预训练模型的状态之间的关系示意图。
图4示出根据本公开实施例的一种预训练模型发布装置的框图。
图5示出根据本公开实施例的一种电子设备的框图。
图6示出根据本公开实施例的一种电子设备的框图。
具体实施方式
以下将参考附图详细说明本公开的各种示例性实施例、特征和方面。附图中相同的附图标记表示功能相同或相似的元件。尽管在附图中示出了实施例的各种方面，但是除非特别指出，不必按比例绘制附图。
在这里专用的词“示例性”意为“用作例子、实施例或说明性”。这里作为“示例性”所说明的任何实施例不必解释为优于或好于其它实施例。
本文中术语“和/或”，是一种描述关联对象的关联关系，表示可以存在三种关系，例如，A和/或B，可以表示：单独存在A，同时存在A和B，单独存在B这三种情况。另外，本文中术语“至少一种”表示多种中的任意一种或多种中的至少两种的任意组合，例如，包括A、B、C中的至少一种，可以表示包括从A、B和C构成的集合中选择的任意一个或多个元素。
另外，为了更好地说明本公开，在下文的具体实施方式中给出了众多的具体细节。本领域技术人员应当理解，没有某些具体细节，本公开同样可以实施。在一些实例中，对于本领域技术人员熟知的方法、手段、元件和电路未作详细描述，以便于凸显本公开的主旨。
在相关技术中，在公司内外部，不同项目组、不同业务线对模型定义不同，可能会导致预训练模型的使用率不高；同时对预训练模型没有统一的管理规范，不同用户预训 练得到的预训练模型没有统一的规范和标准，不便于其它用户对预训练模型的使用，导致预训练模型的质量和可用性不高。
本公开实施例提供的预训练模型发布方法，对预训练模型进行统一的、规范化的集中管理，用户可下载调用已发布的预训练模型进行测试、调优。
在本公开实施例中，通过接收目标对象上传的预训练模型，显示模型信息输入界面；接收所述目标对象在所述输入界面输入的所述预训练模型的模型信息；根据预设的标准对所述模型信息进行核验；在所述模型信息符合所述标准的情况下，对所述预训练模型进行发布。由此，提供了规范化的预训练模型发布流程，并通过预设的标准对模型信息进行核验，使得发布的预训练模型符合预设的标准，减少了因模型信息的不一致导致的适配成本，提高了发布的预训练模型的可用性和便利性。
在一些实施例中，所述预训练模型发布方法的执行主体可以是预训练模型管理平台或电子设备，该平台或电子设备可运行于终端设备或服务器等电子设备中。预训练模型发布方法也可以由终端设备或服务器等电子设备执行，终端设备可以为用户设备(User Equipment，UE)、移动设备、用户终端、终端、蜂窝电话、无绳电话、个人数字助理(Personal Digital Assistant，PDA)、手持设备、计算设备、车载设备、可穿戴设备等，所述方法可以通过处理器调用存储器中存储的计算机可读指令的方式来实现。
为便于描述，在本说明书一个或多个实施例中，预训练模型发布方法的执行主体可以是预训练模型管理平台，后文以执行主体为预训练模型管理平台为例，对该方法的实施方式进行介绍。可以理解，该方法的执行主体为预训练模型管理平台只是一种示例性的说明，并不应理解为对该方法的限定。
图1示出本公开实施例提供的预训练模型发布方法的流程图，如图1所示，所述预训练模型发布方法包括：
在步骤S11中，接收目标对象上传的预训练模型，显示模型信息输入界面；
在一些实施例中，目标对象可以是发明预训练模型的用户。目标对象可以在预训练模型管理平台中上传预训练模型，目标对象上传的预训练模型可以存储至存储器中，以便后续发布。
目标对象上传的预训练模型可以是实现任意功能的模型，例如，可以是用于进行图像识别的模型，也可以是用于进行语音识别的模型，也可以是用于进行文字识别的模型，本公开对目标对象上传的预训练模型的功能不做具体限定。
模型信息输入界面用于供目标对象输入预训练模型的模型信息，模型信息包括用于描述预训练模型的信息，可参见本公开实施例提供的可能的实现方式。
目标对象点击上传模型按钮后，出现上传模型的页面，供目标对象上传预训练模型，输入模型信息。
在步骤S12中，接收所述目标对象在所述输入界面输入的所述预训练模型的模型信息；
在向目标对象显示模型信息输入界面后，目标对象可以在界面中输入预训练模型的模型信息，并提交至预训练模型管理平台，管理平台即可接收到目标对象输入的所述预训练模型的模型信息。
在步骤S13中，根据预设的标准对所述模型信息进行核验；
预设的标准可以是预先设置的对模型信息进行规范的标准，预设的标准可以包括模型信息中包含全部必填信息、模型信息中的模型训练代码存储的仓库地址能够被访问、模型信息中的配置文件和模型的结构文件能够正常可视化、或预训练模型的训练性能指标正常。
在进行核验的过程中，可以逐一核验预训练模型的各模型信息是否符合预设的标准， 在存在一条模型信息不符合预设的标注的情况下，可以向目标对象发送模型信息核验不通过的通知，并向目标对象提供模型信息的修改界面，以便目标对象对预训练模型进行修改。
在步骤S14中，在所述模型信息符合所述标准的情况下，对所述预训练模型进行发布。
预训练模型发布时，目标对象可以填写发布的预训练模型的版本号，确认发布后，模型将会被公开，在对预训练模型进行发布后，其他用户便可以对预训练模型进行下载、测试或调优。
在一些实施例中，对预训练模型进行发布可以是将预训练模型发布至指定的用户可访问的平台，对预训练模型进行发布也可以是将预训练模型设置为指定的用户可访问的状态，这里的指定的用户可以是部分用户，也可以是任意用户。
在本公开实施例中，通过接收目标对象上传的预训练模型和预训练模型的模型信息，根据预设的标准对模型信息进行核验，在模型信息符合标准的情况下，对预训练模型进行发布。由此，本公开实施例提供了规范化的预训练模型发布流程，并通过预设的标准对模型信息进行核验，使得发布的预训练模型符合预设的标准，减少了因模型信息的不一致导致的适配成本，提高了发布的预训练模型的可用性和便利性。
在一些实施例中，所述模型信息包括下述至少一种：模型名称、模型的训练任务类别、模型训练代码存储的仓库地址、模型所属的业务项目、模型占用存储空间的大小、模型的配置文件、模型的结构文件、模型的训练性能指标。
请参阅图2，图2为本公开提供的一种模型信息输入界面的示意图。
其中，模型名称(name)，用来命名预训练模型，模型名称是必填项。
模型的训练任务类别(task)，用于表征模型的训练任务所属的类别，在一些实施例中，该类别可以是检测任务类别，检测任务类别为执行目标检测任务的模型所属的类别，例如，用于在图像中检测人脸的模型中，或者用于在图像中检测猫的模型中。该类别还可以是分类任务类别，分类任务类别为执行分类任务的模型所属的类别，例如，用于对图像进行分类的模型中。
模型训练代码存储的仓库地址(git repo)，用来指示模型训练程序在网络中的存储位置，模型训练程序是用来对预训练模型进行训练的程序，该程序通过训练样本对预训练模型进行训练。
模型所属的业务项目(tag)，可以是企业中的某一业务项目，即预训练模型是属于企业中的哪一业务项目。
模型占用存储空间的大小(model structure)，用于表征模型占用的存储器的存储空间，在一些实施例中，模型占用存储空间的大小可以是占用存储空间的具体值，例如：5G；模型占用存储空间的大小也可以是用来粗略表示存储空间大小的词，例如可以是小、中等、大，等等。
模型的配置文件(config)中包含模型参数的配置信息。
模型的结构文件(weights)中包含用于表征模型结构的参数，例如可以是模型中节点的权重参数。
模型的训练性能指标(metric)，用于表征模型在某个训练集上的性能，例如，可以是模型在各个样本集上的准确度等等。
此外，模型信息中还可以包括其它的附加信息。
在本公开实施例中，模型训练代码存储的仓库地址、模型的配置文件、模型的结构文件这三个字段能够唯一定义一个预训练模型，因此在更改模型信息时，不允许修改这三个字段。
在本公开实施例中，可以通过统一预训练模型的模型信息，使得预训练模型的使用者可以更好地去适配模型，尤其是针对公司内部而言，降低了公司内部因模型信息的不一致，而导致的适配成本。
在一些实施例中，所述预设的标准包括下述至少一种：所述模型信息中包含所有必填信息；所述模型信息中的模型训练代码存储的仓库地址能够被访问；所述模型信息中的配置文件和模型的结构文件能够正常可视化；所述预训练模型的训练性能指标正常。
在本公开实施例中，可以规定模型信息中的必填信息，在一些实施例中，必填信息可以是模型名称、模型的训练任务类别、模型训练代码存储的仓库地址、模型所属的业务项目、模型的配置文件、模型的结构文件。
在一些实施例中，还可以对模型信息中的模型训练代码存储的仓库地址进行访问检测，以确定其是否能够被访问；还可以核验所述模型信息中的配置文件和模型的结构文件是否能够正常可视化，以便正确地查看和编辑配置文件和结构文件；还可以核验预训练模型的训练性能指标是否正常。在一些实施例中，当确定性能指标中的准确率大于100％的情况下，则可以确定模型的训练性能指标不正常，核验不通过。
在本公开实施例中，通过预设的标准来对模型信息进行核验，使得发布的预训练模型符合预设的标准，减少了因模型信息的不一致导致的适配成本，提高了发布的预训练模型的可用性和便利性。用户通过查看模型信息，能够确定模型的性能效果，以决定是否使用该模型；而模型训练代码存储的仓库地址能够正常访问，使得使用模型的用户可以正常调用这个模型，用于之后的模型调优。
在一些实施例中，所述接收目标对象上传的预训练模型，包括：接收目标对象上传至后台管理平台的预训练模型；所述对所述预训练模型进行发布，包括：将所述预训练模型发布至预设平台。
在本公开实施例中，预设平台可以是任一公开的、用户可访问的平台。预训练模型可以分两份数据，在预训练模型处于未发布的状态下，预训练模型存储于后台管理平台中，在后台管理平台中的预训练模型可供上传用户本人可见，也可供后台管理平台中的管理员用户可见；而在预训练模型处于已发布的状态下，预训练模型会复制一份至公开平台中进行存储，以提高数据的安全性，在公开平台中的预训练模型可供用户进行下载、测试、调优。
在一些实施例中，所述方法还包括：接收目标对象上传的用于对已发布的所述预训练模型进行更新的更新模型，向目标对象显示模型信息输入界面；接收目标对象输入的所述更新模型的模型信息；根据预设的标准对所述更新模型的模型信息进行核验，得到核验结果；根据所述核验结果，确定所述更新模型的模型信息在符合所述标准的情况下，对所述更新模型进行发布。
在一些实施例中，可以支持目标对象对已经发布的预训练模型进行更新，其更新过程与上传新的预训练模型的过程类似，也可以向目标对象显示模型信息输入界面，该输入界面供目标对象输入更新模型的模型信息，更新模型的模型信息可参照前文模型信息的相关描述；目标对象提交模型信息后，即可接收目标对象输入的更新模型的模型信息，然后根据预设的标准对更新模型的模型信息进行核验，得到核验结果的核验过程可参照前文对模型信息核验的相关描述；根据所述核验结果，确定所述更新模型的模型信息在符合所述标准的情况下，对更新模型进行发布。
在对更新模型进行发布时，可以是将更新模型发布至公开平台，用户便可以对更新模型进行下载、测试、调优。
在本公开实施例中，更新模型同样可以分两份数据，在更新模型处于未发布的状态下，更新模型存储于后台管理平台中，可供上传用户本人可见，也可供后台管理平台中 的管理员用户可见；而公开平台中的预训练模型为之前公开的预训练模型，在更新模型处于已发布的状态下，更新模型会复制一份至公开平台中进行存储，以提高数据的安全性，在公开平台中的更新模型可供用户进行下载、测试、调优。
在一些实施例中，预训练模型发布方法还包括：对所述更新模型和所述预训练模型进行比对，确定更新模型和所述预训练模型之间的更新信息，根据所述更新信息生成预训练模型版本更新说明。
本公开实施例在对已发布的预训练模型的版本进行更新后，可以在界面中的每一区域显示此版更新的更新信息(例如发版通知)，即该版本相对于上一版本所更新的信息，同时，也可以显示该版本的模型信息。
在本公开实施例中，可以通过对更新模型和预训练模型进行比对，来确定更新模型和预训练模型之间的更新信息，即确定更新模型和预训练模型之间的差异部分。在确定更新信息后，即可将该更新信息作为对预训练模型的版本更新说明进行显示。
在本公开实施例中，通过对所述更新模型和所述预训练模型进行比对，确定更新模型和所述预训练模型之间的更新信息；根据所述更新信息生成预训练模型版本更新说明。由此，能够自动生成更新模型的更新说明，并进行显示，提高了发版通知的生成效率。
在一些实施例中，所述预训练模型的状态包括下述至少一种：未发布状态、已发布状态、已更新状态、已下线状态；所述未发布状态的预训练模型支持的操作包括：发布、编辑、查看详情、删除；所述已发布状态的预训练模型支持的操作包括：编辑、查看详情、取消发布；所述已更新状态的预训练模型支持的操作包括：发布、编辑、查看详情、取消发布；所述已下线状态的预训练模型支持的操作包括：删除。
请参阅图3，图3为本公开提供的一种预训练模型的状态之间的关系示意图，在本公开实施例中，目标对象在将预训练模型上传后，预训练模型处于未发布状态301，此时目标对象可以对预训练模型执行发布、编辑、查看详情、删除等操作；在预训练模型发布后，预训练模型处于已发布状态302，目标对象可以对预训练模型执行编辑、查看详情、取消发布等操作；在目标对象对已发布状态302的预训练模型进行取消发布后，预训练模型会进入已下线状态303，针对已下线状态303的预训练模型，目标对象可以对预训练模型执行删除操作；在目标对象对已发布状态302的预训练模型进行更新并发布后，更新模型进入已更新状态304，处于已更新状态304的模型可以支持目标对象执行发布、编辑、查看详情、取消发布等操作。
可以理解，本公开提及的上述各个方法实施例，在不违背原理逻辑的情况下，均可以彼此相互结合形成结合后的实施例。本领域技术人员可以理解，在具体实施方式的上述方法中，各步骤的具体执行顺序应当以其功能和可能的内在逻辑确定。
此外，本公开实施例还提供了预训练模型发布装置、电子设备、计算机可读存储介质核计算机程序产品，均可用来实现本公开实施例提供的任一种预训练模型发布方法，相应技术方案和描述和参见方法部分的相应记载。
图4示出本公开实施例提供的预训练模型发布装置的框图，如图4所示，所述装置20包括：
第一上传部分21，配置为接收目标对象上传的预训练模型，显示模型信息输入界面；
第一信息接收部分22，配置为接收所述目标对象在所述输入界面输入的所述预训练模型的模型信息；
第一核验部分23，配置为根据预设的标准对所述模型信息进行核验；
第一发布部分24，配置为在所述模型信息符合所述标准的情况下，对所述预训练模型进行发布。
在一些实施例中，所述第一上传部分，还配置为接收目标对象上传至后台管理平台 的预训练模型；所述第一发布部分，还配置为将所述预训练模型发布至预设平台。
在一些实施例中，所述模型信息包括下述至少一种：
模型名称、模型的训练任务类别、模型训练代码存储的仓库地址、模型所属的业务项目、模型占用存储空间的大小、模型的配置文件、模型的结构文件、模型的训练性能指标。
在一些实施例中，所述预设的标准包括下述至少一种：
所述模型信息中包含所有必填信息；
所述模型信息中的模型训练代码存储的仓库地址能够被访问；
所述模型信息中的配置文件和模型的结构文件能够正常可视化；
所述预训练模型的训练性能指标正常。
在一些实施例中，所述装置还包括：
第二上传部分，配置为接收目标对象上传的配置为对已发布的所述预训练模型进行更新的更新模型，显示所述更新模型的模型信息输入界面；
第二信息接收部分，配置为接收目标对象输入的所述更新模型的模型信息；
第二核验部分，配置为根据预设的标准对所述更新模型的模型信息进行核验，得到核验结果；
第二发布部分，配置为根据所述核验结果，确定所述更新模型的模型信息在符合所述标准的情况下，对所述更新模型进行发布。
在一些实施例中，所述装置还包括：
比对部分，配置为对所述更新模型和所述预训练模型进行比对，确定更新模型和所述预训练模型之间的更新信息；
更新说明生成部分，配置为根据所述更新信息生成预训练模型版本更新说明。
在一些实施例中，所述预训练模型的状态包括下述至少一种：
未发布状态、已发布状态、已更新状态、已下线状态；
所述未发布状态的预训练模型支持的操作包括：发布、编辑、查看详情、删除；
所述已发布状态的预训练模型支持的操作包括：编辑、查看详情、取消发布；
所述已更新状态的预训练模型支持的操作包括：发布、编辑、查看详情、取消发布；
所述已下线状态的预训练模型支持的操作包括：删除。
在一些实施例中，本公开实施例提供的装置具有的功能或包含的部分可以配置为执行上文方法实施例描述的方法，其具体实现可以参照上文方法实施例的描述。
在本公开实施例以及其他的实施例中，“部分”可以是部分电路、部分处理器、部分程序或软件等等，当然也可以是单元，还可以是模块也可以是非模块化的。
本公开实施例还提出一种计算机可读存储介质，其上存储有计算机程序指令，所述计算机程序指令被处理器执行时实现上述方法。计算机可读存储介质可以是易失性或非易失性计算机可读存储介质。
本公开实施例还提出一种电子设备，包括：处理器；配置为存储处理器可执行指令的存储器；其中，所述处理器配置为调用所述存储器存储的指令，以执行上述方法。
本公开实施例还提供了一种计算机程序产品，包括计算机可读代码，或者承载有计算机可读代码的非易失性计算机可读存储介质，当所述计算机可读代码在电子设备的处理器中运行时，所述电子设备中的处理器执行上述方法。
在本公开实施例中，电子设备可以被提供为终端、服务器或其它形态的设备。
图5示出本公开实施例提供的一种电子设备800的框图。例如，电子设备800可以是用户设备(User Equipment，UE)、移动设备、用户终端、终端、蜂窝电话、无绳电话、个人数字处理(Personal Digital Assistant，PDA)、手持设备、计算设备、车载设备、 可穿戴设备等终端设备。
参照图5，电子设备800可以包括以下一个或多个组件：处理组件802，存储器804，电源组件806，多媒体组件808，音频组件810，输入/输出(Input/Output，I/O)接口812，传感器组件814，以及通信组件816。
处理组件802通常控制电子设备800的整体操作，诸如与显示、电话呼叫、数据通信、相机操作和记录操作相关联的操作。处理组件802可以包括一个或多个处理器820来执行指令，以完成上述的方法的全部或部分步骤。此外，处理组件802可以包括一个或多个部分，便于处理组件802和其他组件之间的交互。例如，处理组件802可以包括多媒体部分，以方便多媒体组件808和处理组件802之间的交互。
存储器804配置为存储各种类型的数据以支持在电子设备800的操作。这些数据的示例包括配置为在电子设备800上操作的任何应用程序或方法的指令，联系人数据，电话簿数据，消息，图片，视频等。存储器804可以由任何类型的易失性或非易失性存储设备或者它们的组合实现，如静态随机存取存储器(Static Random-Access Memory，SRAM)，电可擦除可编程只读存储器(Electrically Erasable Programmable Read Only Memory，EEPROM)，可擦除可编程只读存储器(Erasable Programmable Read Only Memory，EPROM)，可编程只读存储器(Programmable Read Only Memory，PROM)，只读存储器(Read Only Memory，ROM)，磁存储器，快闪存储器，磁盘或光盘。
电源组件806为电子设备800的各种组件提供电力。电源组件806可以包括电源管理系统，一个或多个电源，及其他与为电子设备800生成、管理和分配电力相关联的组件。
多媒体组件808包括在所述电子设备800和用户之间的提供一个输出接口的屏幕。在一些实施例中，屏幕可以包括液晶显示器(Liquid Crystal Display，LCD)和触摸面板(Touch Panel，TP)。如果屏幕包括触摸面板，屏幕可以被实现为触摸屏，以接收来自用户的输入信号。触摸面板包括一个或多个触摸传感器以感测触摸、滑动和触摸面板上的手势。所述触摸传感器可以不仅感测触摸或滑动动作的边界，而且还检测与所述触摸或滑动操作相关的持续时间和压力。在一些实施例中，多媒体组件808包括一个前置摄像头和/或后置摄像头。当电子设备800处于操作模式，如拍摄模式或视频模式时，前置摄像头和/或后置摄像头可以接收外部的多媒体数据。每个前置摄像头和后置摄像头可以是一个固定的光学透镜系统或具有焦距和光学变焦能力。
音频组件810配置为输出和/或输入音频信号。例如，音频组件810包括一个麦克风(Microphone，MIC)，当电子设备800处于操作模式，如呼叫模式、记录模式和语音识别模式时，麦克风配置为接收外部音频信号。所接收的音频信号可以被进一步存储在存储器804或经由通信组件816发送。在一些实施例中，音频组件810还包括一个扬声器，配置为输出音频信号。
I/O接口812为处理组件802和外围接口部分之间提供接口，上述外围接口部分可以是键盘，点击轮，按钮等。这些按钮可包括但不限于：主页按钮、音量按钮、启动按钮和锁定按钮。
传感器组件814包括一个或多个传感器，配置为电子设备800提供各个方面的状态评估。例如，传感器组件814可以检测到电子设备800的打开或关闭状态，组件的相对定位，例如所述组件为电子设备800的显示器和小键盘，传感器组件814还可以检测电子设备800或电子设备800中组件的位置改变，用户与电子设备800接触的存在或不存在，电子设备800方位或加速或减速和电子设备800的温度变化。传感器组件814可以包括接近传感器，被配置用来在没有任何的物理接触时检测附近物体的存在。传感器组件814还可以包括光传感器，如互补金属氧化物半导体(Complementary Metal Oxide  Semiconductor Sensor，CMOS)或电荷耦合装置(Charge-Coupled Device，CCD)图像传感器，配置为在成像应用中使用。在一些实施例中，该传感器组件814还可以包括加速度传感器，陀螺仪传感器，磁传感器，压力传感器或温度传感器。
通信组件816配置为便于电子设备800和其他设备之间有线或无线方式的通信。电子设备800可以接入基于通信标准的无线网络，如无线网络(Wireless Fidelity，Wi-Fi)、第二代移动通信技术(2nd Generation，2G)、第三代移动通信技术(3rd Generation，3G)、第四代移动通信技术(4th Generation，4G)、通用移动通信技术的长期演进(Long Term Evolution，LTE)、第五代移动通信技术(5th Generation，5G)或它们的组合。在一个示例性实施例中，通信组件816经由广播信道接收来自外部广播管理系统的广播信号或广播相关信息。在一个示例性实施例中，所述通信组件816还包括近场通信(Near Field Communication，NFC)部分，以促进短程通信。例如，在NFC部分可基于射频识别(Radio Frequency Identification，RFID)技术，红外数据协会(Infrared-Data-Association，IrDA)技术，超宽带(Ultra Wide Band，UWB)技术，蓝牙(Bluetooth，BT)技术和其他技术来实现。
在示例性实施例中，电子设备800可以被一个或多个应用专用集成电路(Application-Specific Integrated Circuit，ASIC)、数字信号处理器(Digital System Processor，DSP)、数字信号处理设备(Digital Signal Processing Devices，DSPD)、可编程逻辑器件(Programmable Logic Device，PLD)、现场可编程门阵列(Field Programmable Gate Array，FPGA)、控制器、微控制器、微处理器或其他电子元件实现，配置为执行上述方法。
在示例性实施例中，还提供了一种非易失性计算机可读存储介质，例如包括计算机程序指令的存储器804，上述计算机程序指令可由电子设备800的处理器820执行以完成上述方法。
本公开实施例涉及增强现实领域，通过获取现实环境中的目标对象的图像信息，进而借助各类视觉相关算法实现对目标对象的相关特征、状态及属性进行检测或识别处理，从而得到与具体应用匹配的虚拟与现实相结合的AR效果。示例性的，目标对象可涉及与人体相关的脸部、肢体、手势、动作等，或者与物体相关的标识物、标志物，或者与场馆或场所相关的沙盘、展示区域或展示物品等。视觉相关算法可涉及视觉定位、即时定位与地图构建(Simultaneous Localization and Mapping，SLAM)、三维重建、图像注册、背景分割、对象的关键点提取及跟踪、对象的位姿或深度检测等。具体应用不仅可以涉及跟真实场景或物品相关的导览、导航、讲解、重建、虚拟效果叠加展示等交互场景，还可以涉及与人相关的特效处理，比如妆容美化、肢体美化、特效展示、虚拟模型展示等交互场景。可通过卷积神经网络，实现对目标对象的相关特征、状态及属性进行检测或识别处理。上述卷积神经网络是基于深度学习框架进行模型训练而得到的网络模型。
图6示出根据本公开实施例的一种电子设备1900的框图。例如，电子设备1900可以被提供为一服务器。参照图6，电子设备1900包括处理组件1922，其进一步包括一个或多个处理器，以及由存储器1932所代表的存储器资源，用于存储可由处理组件1922的执行的指令，例如应用程序。存储器1932中存储的应用程序可以包括一个或一个以上的每一个对应于一组指令的部分。此外，处理组件1922配置为执行指令，以执行上述方法。
电子设备1900还可以包括一个电源组件1926配置为执行电子设备1900的电源管理，一个有线或无线网络接口1950配置为将电子设备1900连接到网络，和一个输入输出(I/O)接口1958。电子设备1900可以操作基于存储在存储器1932的操作系统，例 如微软服务器操作系统(Windows Server TM)，苹果公司推出的基于图形用户界面操作系统(Mac OS XTM)，多用户多进程的计算机操作系统(Unix TM)，自由和开放原代码的类Unix操作系统(Linux TM)，开放原代码的类Unix操作系统(FreeBSD TM)或类似。
在一些实施例中，还提供了一种非易失性计算机可读存储介质，例如包括计算机程序指令的存储器1932，上述计算机程序指令可由电子设备1900的处理组件1922执行以完成上述方法。
本公开实施例可以是系统、方法和/或计算机程序产品。计算机程序产品可以包括计算机可读存储介质，其上载有配置为使处理器实现本公开实施例的各个方面的计算机可读程序指令。
计算机可读存储介质可以是可以保持和存储由指令执行设备使用的指令的有形设备。计算机可读存储介质例如可以是(但不限于)电存储设备、磁存储设备、光存储设备、电磁存储设备、半导体存储设备或者上述的任意合适的组合。计算机可读存储介质的更具体的例子(非穷举的列表)包括：便携式计算机盘、硬盘、随机存取存储器(RAM)、只读存储器(ROM)、可擦式可编程只读存储器(EPROM或闪存)、静态随机存取存储器(SRAM)、便携式压缩盘只读存储器(Compact Disc Read-Only Memory，CD-ROM)、数字多功能盘(Digital Versatile Disc，DVD)、记忆棒、软盘、机械编码设备、例如其上存储有指令的打孔卡或凹槽内凸起结构、以及上述的任意合适的组合。这里所使用的计算机可读存储介质不被解释为瞬时信号本身，诸如无线电波或者其他自由传播的电磁波、通过波导或其他传输媒介传播的电磁波(例如，通过光纤电缆的光脉冲)、或者通过电线传输的电信号。
这里所描述的计算机可读程序指令可以从计算机可读存储介质下载到各个计算/处理设备，或者通过网络、例如因特网、局域网、广域网和/或无线网下载到外部计算机或外部存储设备。网络可以包括铜传输电缆、光纤传输、无线传输、路由器、防火墙、交换机、网关计算机和/或边缘服务器。每个计算/处理设备中的网络适配卡或者网络接口从网络接收计算机可读程序指令，并转发该计算机可读程序指令，以供存储在各个计算/处理设备中的计算机可读存储介质中。
配置为执行本公开实施例操作的计算机程序指令可以是汇编指令、指令集架构(Instruction Set Architecture，ISA)指令、机器指令、机器相关指令、微代码、固件指令、状态设置数据、或者以一种或多种编程语言的任意组合编写的源代码或目标代码，所述编程语言包括面向对象的编程语言—诸如Smalltalk、C++等，以及常规的过程式编程语言—诸如“C”语言或类似的编程语言。计算机可读程序指令可以完全地在用户计算机上执行、部分地在用户计算机上执行、作为一个独立的软件包执行、部分在用户计算机上部分在远程计算机上执行、或者完全在远程计算机或服务器上执行。在涉及远程计算机的情形中，远程计算机可以通过任意种类的网络—包括局域网(Local Area Network，LAN)或广域网(Wide Area Network，WAN)—连接到用户计算机，或者，可以连接到外部计算机(例如利用因特网服务提供商来通过因特网连接)。在一些实施例中，通过利用计算机可读程序指令的状态信息来个性化定制电子电路，例如可编程逻辑电路、现场可编程门阵列(FPGA)或可编程逻辑阵列(Programmable Array Logic，PLA)，该电子电路可以执行计算机可读程序指令，从而实现本公开的各个方面。
这里参照根据本公开实施例的方法、装置(系统)和计算机程序产品的流程图和/或框图描述了本公开的各个方面。应当理解，流程图和/或框图的每个方框以及流程图和/或框图中各方框的组合，都可以由计算机可读程序指令实现。
这些计算机可读程序指令可以提供给通用计算机、专用计算机或其它可编程数据处 理装置的处理器，从而生产出一种机器，使得这些指令在通过计算机或其它可编程数据处理装置的处理器执行时，产生了实现流程图和/或框图中的一个或多个方框中规定的功能/动作的装置。也可以把这些计算机可读程序指令存储在计算机可读存储介质中，这些指令使得计算机、可编程数据处理装置和/或其他设备以特定方式工作，从而，存储有指令的计算机可读介质则包括一个制造品，其包括实现流程图和/或框图中的一个或多个方框中规定的功能/动作的各个方面的指令。
也可以把计算机可读程序指令加载到计算机、其它可编程数据处理装置、或其它设备上，使得在计算机、其它可编程数据处理装置或其它设备上执行一系列操作步骤，以产生计算机实现的过程，从而使得在计算机、其它可编程数据处理装置、或其它设备上执行的指令实现流程图和/或框图中的一个或多个方框中规定的功能/动作。
附图中的流程图和框图显示了根据本公开的多个实施例的系统、方法和计算机程序产品的可能实现的体系架构、功能和操作。在这点上，流程图或框图中的每个方框可以代表一个部分、程序段或指令的一部分，所述部分、程序段或指令的一部分包含一个或多个配置为实现规定的逻辑功能的可执行指令。在有些作为替换的实现中，方框中所标注的功能也可以以不同于附图中所标注的顺序发生。例如，两个连续的方框实际上可以基本并行地执行，它们有时也可以按相反的顺序执行，这依所涉及的功能而定。也要注意的是，框图和/或流程图中的每个方框、以及框图和/或流程图中的方框的组合，可以用执行规定的功能或动作的专用的基于硬件的系统来实现，或者可以用专用硬件与计算机指令的组合来实现。
该计算机程序产品可以具体通过硬件、软件或其结合的方式实现。在一个可选实施例中，所述计算机程序产品具体体现为计算机存储介质，在另一个可选实施例中，计算机程序产品具体体现为软件产品，例如软件开发包(Software Development Kit，SDK)等等。
以上已经描述了本公开的各实施例，上述说明是示例性的，并非穷尽性的，并且也不限于所披露的各实施例。在不偏离所说明的各实施例的范围和精神的情况下，对于本技术领域的普通技术人员来说许多修改和变更都是显而易见的。本文中所用术语的选择，旨在最好地解释各实施例的原理、实际应用或对市场中的技术的改进，或者使本技术领域的其它普通技术人员能理解本文披露的各实施例。
工业实用性
本公开实施例提供了一种预训练模型发布方法及装置、电子设备和存储介质和计算机程序产品，接收目标对象上传的预训练模型，显示模型信息输入界面；接收所述目标对象在所述输入界面输入的所述预训练模型的模型信息；根据预设的标准对所述模型信息进行核验；在所述模型信息符合所述标准的情况下，对所述预训练模型进行发布。由此，提供了规范化的预训练模型发布流程，并通过预设的标准对模型信息进行核验，使得发布的预训练模型符合预设的标准，减少了因模型信息的不一致导致的适配成本，提高了发布的预训练模型的可用性和便利性。

Claims (17)

  1. 一种预训练模型发布方法，所述方法应用于电子设备，所述方法包括：
    接收目标对象上传的预训练模型，显示模型信息输入界面；
    接收所述目标对象在所述输入界面输入的所述预训练模型的模型信息；
    根据预设的标准对所述模型信息进行核验；
    在所述模型信息符合所述标准的情况下，对所述预训练模型进行发布。
  2. 根据权利要求1所述的方法，其中，所述接收目标对象上传的预训练模型，包括：接收目标对象上传至后台管理平台的所述预训练模型；
    所述对所述预训练模型进行发布，包括：将所述预训练模型发布至预设平台。
  3. 根据权利要求1或2所述的方法，其中，所述模型信息包括下述至少一种：
    模型名称、模型的训练任务类别、模型训练代码存储的仓库地址、模型所属的业务项目、模型占用存储空间的大小、模型的配置文件、模型的结构文件、模型的训练性能指标。
  4. 根据权利要求3所述的方法，其中，所述预设的标准包括下述至少一种：
    所述模型信息中包含所有必填信息；
    所述模型信息中的模型训练代码存储的仓库地址能够被访问；
    所述模型信息中的配置文件和模型的结构文件能够正常可视化；
    所述预训练模型的训练性能指标正常。
  5. 根据权利要求1-4任一所述的方法，其中，所述方法还包括：
    接收目标对象上传的用于对已发布的所述预训练模型进行更新的更新模型，显示所述更新模型的模型信息输入界面；
    接收目标对象输入的所述更新模型的模型信息；
    根据预设的标准对所述更新模型的模型信息进行核验，得到核验结果；
    根据所述核验结果，确定所述更新模型的模型信息在符合所述标准的情况下，对所述更新模型进行发布。
  6. 根据权利要求5所述的方法，其中，所述方法还包括：
    对所述更新模型和所述预训练模型进行比对，确定更新模型和所述预训练模型之间的更新信息；
    根据所述更新信息生成预训练模型版本更新说明。
  7. 根据权利要求1-6任一所述的方法，其中，所述预训练模型的状态包括下述至少一种：
    未发布状态、已发布状态、已更新状态、已下线状态；
    所述未发布状态的预训练模型支持的操作包括：发布、编辑、查看详情、删除；
    所述已发布状态的预训练模型支持的操作包括：编辑、查看详情、取消发布；
    所述已更新状态的预训练模型支持的操作包括：发布、编辑、查看详情、取消发布；
    所述已下线状态的预训练模型支持的操作包括：删除。
  8. 一种预训练模型发布装置，所述装置包括：
    第一上传部分，配置为接收目标对象上传的预训练模型，显示模型信息输入界面；
    第一信息接收部分，配置为接收所述目标对象在所述输入界面输入的所述预训练模型的模型信息；
    第一核验部分，配置为根据预设的标准对所述模型信息进行核验；
    第一发布部分，配置为在所述模型信息符合所述标准的情况下，对所述预训练模型 进行发布。
  9. 根据权利要求8所述的装置，其中，所述第一上传部分，还配置为接收目标对象上传至后台管理平台的所述预训练模型；所述第一发布部分，还配置为将所述预训练模型发布至预设平台。
  10. 根据权利要求8或9所述的装置，其中，所述模型信息包括下述至少一种：
    模型名称、模型的训练任务类别、模型训练代码存储的仓库地址、模型所属的业务项目、模型占用存储空间的大小、模型的配置文件、模型的结构文件、模型的训练性能指标。
  11. 根据权利要求8-10任一所述的装置，其中，所述预设的标准包括下述至少一种：
    所述模型信息中包含所有必填信息；
    所述模型信息中的模型训练代码存储的仓库地址能够被访问；
    所述模型信息中的配置文件和模型的结构文件能够正常可视化；
    所述预训练模型的训练性能指标正常。
  12. 根据权利要求8-11任一所述的装置，其中，所述装置还包括：
    第二上传部分，配置为接收目标对象上传的配置为对已发布的所述预训练模型进行更新的更新模型，显示所述更新模型的模型信息输入界面；
    第二信息接收部分，配置为接收目标对象输入的所述更新模型的模型信息；
    第二核验部分，配置为根据预设的标准对所述更新模型的模型信息进行核验，得到核验结果；
    第二发布部分，配置为根据所述核验结果，确定所述更新模型的模型信息在符合所述标准的情况下，对所述更新模型进行发布。
  13. 根据权利要求12所述的装置，其中，所述装置还包括：
    比对部分，配置为对所述更新模型和所述预训练模型进行比对，确定更新模型和所述预训练模型之间的更新信息；
    更新说明生成部分，配置为根据所述更新信息生成预训练模型版本更新说明。
  14. 根据权利要求8-13任一所述的装置，其中，所述预训练模型的状态包括下述至少一种：
    未发布状态、已发布状态、已更新状态、已下线状态；
    所述未发布状态的预训练模型支持的操作包括：发布、编辑、查看详情、删除；
    所述已发布状态的预训练模型支持的操作包括：编辑、查看详情、取消发布；
    所述已更新状态的预训练模型支持的操作包括：发布、编辑、查看详情、取消发布；
    所述已下线状态的预训练模型支持的操作包括：删除。
  15. 一种电子设备，所述电子设备包括：
    处理器；
    配置为存储处理器可执行指令的存储器；
    其中，所述处理器配置为调用所述存储器存储的指令，所述处理器执行权利要求1至7中任意一项所述的方法。
  16. 一种计算机可读存储介质，其上存储有计算机程序指令，所述计算机程序指令被处理器执行时实现权利要求1至7中任意一项所述的方法。
  17. 一种计算机程序产品，所述计算机程序产品包括计算机程序或指令，在所述计算机程序或指令在电子设备上运行的情况下，使得所述电子设备执行权利要求1至7中任意一项所述的方法。
PCT/CN2022/088585 2021-11-30 2022-04-22 预训练模型发布方法及装置、电子设备、存储介质和计算机程序产品 WO2023097952A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202111440634.5 2021-11-30
CN202111440634.5A CN114153540A (zh) 2021-11-30 2021-11-30 预训练模型发布方法及装置、电子设备和存储介质

Publications (1)

Publication Number Publication Date
WO2023097952A1 true WO2023097952A1 (zh) 2023-06-08

Family

ID=80455054

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/088585 WO2023097952A1 (zh) 2021-11-30 2022-04-22 预训练模型发布方法及装置、电子设备、存储介质和计算机程序产品

Country Status (2)

Country Link
CN (1) CN114153540A (zh)
WO (1) WO2023097952A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114153540A (zh) * 2021-11-30 2022-03-08 上海商汤科技开发有限公司 预训练模型发布方法及装置、电子设备和存储介质

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109271602A (zh) * 2018-09-05 2019-01-25 腾讯科技（深圳）有限公司 深度学习模型发布方法及装置
US20200242508A1 (en) * 2019-01-30 2020-07-30 Open Text Sa Ulc Machine learning model publishing systems and methods
CN111612158A (zh) * 2020-05-22 2020-09-01 云知声智能科技股份有限公司 模型部署方法、装置、设备和存储介质
CN112784778A (zh) * 2021-01-28 2021-05-11 北京百度网讯科技有限公司 生成模型并识别年龄和性别的方法、装置、设备和介质
CN113052328A (zh) * 2021-04-02 2021-06-29 上海商汤科技开发有限公司 深度学习模型生产系统、电子设备和存储介质
CN114153540A (zh) * 2021-11-30 2022-03-08 上海商汤科技开发有限公司 预训练模型发布方法及装置、电子设备和存储介质

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11720813B2 (en) * 2017-09-29 2023-08-08 Oracle International Corporation Machine learning platform for dynamic model selection
US11443239B2 (en) * 2020-03-17 2022-09-13 Microsoft Technology Licensing, Llc Interface for machine teaching modeling
CN113010441B (zh) * 2021-04-29 2024-05-07 成都新希望金融信息有限公司 模型发布方法、装置、电子设备及存储介质
CN113626116B (zh) * 2021-07-20 2023-12-15 中国电子科技集团公司电子科学研究院 智能学习系统及数据分析方法

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109271602A (zh) * 2018-09-05 2019-01-25 腾讯科技（深圳）有限公司 深度学习模型发布方法及装置
US20200242508A1 (en) * 2019-01-30 2020-07-30 Open Text Sa Ulc Machine learning model publishing systems and methods
CN111612158A (zh) * 2020-05-22 2020-09-01 云知声智能科技股份有限公司 模型部署方法、装置、设备和存储介质
CN112784778A (zh) * 2021-01-28 2021-05-11 北京百度网讯科技有限公司 生成模型并识别年龄和性别的方法、装置、设备和介质
CN113052328A (zh) * 2021-04-02 2021-06-29 上海商汤科技开发有限公司 深度学习模型生产系统、电子设备和存储介质
CN114153540A (zh) * 2021-11-30 2022-03-08 上海商汤科技开发有限公司 预训练模型发布方法及装置、电子设备和存储介质

Also Published As

Publication number Publication date
CN114153540A (zh) 2022-03-08

Similar Documents

Publication Publication Date Title
TWI781359B (zh) 人臉和人手關聯檢測方法及裝置、電子設備和電腦可讀儲存媒體
CN109977847A (zh) 图像生成方法及装置、电子设备和存储介质
TWI757668B (zh) 網路優化方法及裝置、圖像處理方法及裝置、儲存媒體
EP3125164A1 (en) Method and device for presenting ticket information
CN110659690B (zh) 神经网络的构建方法及装置、电子设备和存储介质
WO2022188305A1 (zh) 信息展示方法及装置、电子设备、存储介质及计算机程序
TW202109360A (zh) 圖像處理方法及圖像處理裝置、電子設備和電腦可讀儲存介質
CN114240882A (zh) 缺陷检测方法及装置、电子设备和存储介质
CN112668707B (zh) 运算方法、装置及相关产品
CN110569329B (zh) 数据处理方法及装置、电子设备和存储介质
WO2023123840A1 (zh) 支付方法及装置、电子设备、存储介质和计算机程序产品
CN111242303A (zh) 网络训练方法及装置、图像处理方法及装置
CN112561461A (zh) 基于机器学习的政务审批方法、系统、装置及存储介质
CN111523599B (zh) 目标检测方法及装置、电子设备和存储介质
CN114546460A (zh) 固件升级方法及装置、电子设备和存储介质
WO2023097952A1 (zh) 预训练模型发布方法及装置、电子设备、存储介质和计算机程序产品
WO2023024439A1 (zh) 一种行为识别方法及装置、电子设备和存储介质
CN109447258B (zh) 神经网络模型的优化方法及装置、电子设备和存储介质
CN113807498B (zh) 模型扩展方法及装置、电子设备和存储介质
CN112559673A (zh) 语言处理模型的训练方法及装置、电子设备及存储介质
CN111488964A (zh) 图像处理方法及装置、神经网络训练方法及装置
CN108241438B (zh) 一种输入方法、装置和用于输入的装置
CN117112553A (zh) 基于可配置表结构公式计算的碳统计方法、系统和设备
CN111046780A (zh) 神经网络训练及图像识别方法、装置、设备和存储介质
CN106126104B (zh) 键盘模拟方法和装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22899775

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE