Automation platform for lightweight and performance evaluation of deep learning model, model lightweight method and model performance evaluation method

Abstract

The invention provides an automatic platform for deep learning model lightweight and performance evaluation, a model lightweight method and a model performance evaluation method, and relates to the technical field of model lightweight, the automatic platform comprises a user interface module, an algorithm management module, a task scheduling module, a model evaluation module and a result analysis module; wherein the user interface module is used for uploading a model and a data set, configuring a lightweight task and monitoring a task progress; the algorithm management module is used for managing a pre-integrated algorithm and a user-defined algorithm and providing packaging and deployment functions of the algorithm; the task scheduling module is used for automatically scheduling lightweight tasks and optimizing resource allocation according to task parameters configured by a user; the model evaluation module is used for comparing the effects of different lightweight schemes and providing a performance report; and the result analysis module is used for analyzing a lightweight result and providing a visual report. The user experience and the working efficiency can be remarkably improved.

Description

Technical Field

[0001] The present application relates to the technical field of model lightweighting, and specifically, to an automated platform for lightweighting and performance evaluation of deep learning models, a model lightweighting method, and a model performance evaluation method. Background Art

[0002] With the widespread application of deep learning models in various fields, their size and complexity continue to increase, leading to problems such as high computing resource consumption, high deployment costs, and slow inference speed. Therefore, model lightweighting has become the key to improving model efficiency and adapting to more application scenarios.

[0003] The lightweight products currently available on the market are usually limited to certain types of lightweight operations and lack flexibility and customization capabilities. For users who need to perform complex or specific lightweight tasks, existing solutions cannot meet their needs. In addition, the existing lightweight process often requires a deep technical background and manual adjustments, which increases the user's usage threshold. Summary of the invention

[0004] The embodiments of the present application provide an automated platform for deep learning model lightweighting and performance evaluation, a model lightweighting method, and a model performance evaluation method to solve the problems existing in the prior art.

[0005] Other features and advantages of the present application will become apparent from the following detailed description, or may be learned in part by the practice of the present application.

[0006] According to a first aspect of an embodiment of the present application, there is provided an automated platform for lightweighting and performance evaluation of deep learning models, including: a user interface module, an algorithm management module, a task scheduling module, a model evaluation module, and a result analysis module;

[0007] The user interface module is respectively connected to the task scheduling module and the model evaluation module in communication, the task scheduling module is respectively connected to the algorithm management module and the model evaluation module in communication, and the model evaluation module is also connected to the result analysis module in communication;

[0008] The user interface module is used to upload models and data sets, configure lightweight tasks, and monitor task progress;

[0009] The algorithm management module is used to manage pre-integrated algorithms and user-defined algorithms, and provide algorithm packaging and deployment functions;

[0010] The task scheduling module is used to automatically schedule lightweight tasks and optimize resource allocation according to the task parameters configured by the user;

[0011] The model evaluation module is used to compare the effects of different lightweight solutions and provide performance reports;

[0012] The result analysis module is used to analyze lightweight results and provide a visual report.

[0013] In some embodiments of the present application, based on the aforementioned scheme, the algorithm management module includes an algorithm sub-module, and the algorithm sub-module has a built-in lightweight algorithm, and the lightweight algorithm includes a block pruning algorithm, a Qdrop quantization algorithm and a data-free knowledge distillation algorithm.

[0014] In some embodiments of the present application, based on the aforementioned solution, the model evaluation module has a built-in model performance evaluation tool.

[0015] According to a second aspect of an embodiment of the present application, a model lightweight method implemented based on the automatic dialogue platform described in the first aspect is provided, comprising:

[0016] The user uploads the deep learning model file and corresponding dataset that need to be lightweight through the user interface module;

[0017] The user selects the required lightweight algorithm through the algorithm management module and configures the corresponding parameters according to the selected algorithm;

[0018] After configuring all parameters, start the lightweight task through the user interface module;

[0019] The task scheduling module receives the task request and assigns the task to the lightweight algorithm for optimization processing;

[0020] After processing, the optimized lightweight model is output.

[0021] In some embodiments of the present application, based on the aforementioned solution, during the optimization process, the user monitors the task progress and views the task log through the user interface module.

[0022] According to a third aspect of an embodiment of the present application, a computer-readable storage medium is provided, wherein the storage medium stores computer instructions, and when the computer instructions are executed on a computer, the computer executes the method according to the second aspect.

[0023] According to a fourth aspect of the embodiments of the present application, there is provided an electronic device, including: a memory and a processor;

[0024] The memory is used to store computer instructions;

[0025] The processor is used to call the computer instructions stored in the memory so that the electronic device executes the method as described in the second aspect.

[0026] According to the fifth aspect of the embodiments of the present application, there is provided a model performance evaluation method implemented based on the automatic speech platform as described in the first aspect, including:

[0027] The user uploads the model and test data set to be compared and tested through the user interface module;

[0028] The user configures evaluation parameters and evaluation metrics according to the evaluation needs;

[0029] After the configuration is completed, the evaluation task is started through the user interface module, and the task parameters are passed to the model evaluation module;

[0030] The model evaluation module evaluates the model according to the configured evaluation parameters and evaluation metrics, and generates a detailed evaluation report.

[0031] According to the sixth aspect of the embodiments of the present application, there is provided a computer-readable storage medium, in which computer instructions are stored, and when the computer instructions run on the computer, the computer is enabled to execute the method as described in the fifth aspect.

[0032] According to the seventh aspect of the embodiments of the present application, there is provided an electronic device, including: a memory and a processor;

[0033] The memory is used to store computer instructions;

[0034] The processor is used to call the computer instructions stored in the memory, so that the electronic device executes the method as described in the fifth aspect.

[0035] The technical solution of the present application has the following beneficial effects:

[0036] Reduce the technical threshold: Through the intuitive graphical interface and simplified operation process, even non-professionals can easily perform model lightweighting.

[0037] Improve flexibility: Users can select and configure different lightweighting algorithms according to their own needs, and even upload custom algorithms to achieve personalized optimization.

[0038] Accelerate the development cycle: The automated lightweighting production line and built-in evaluation tools can help users quickly iterate and optimize the model, shortening the development time.

[0039] Improve model performance: Through flexible lightweighting strategies, users can find the most suitable model configuration for their application scenarios, thereby improving the running efficiency and inference speed of the model.

[0040] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present application. Description of the Drawings

[0041] The drawings herein are incorporated into the specification and constitute a part of the specification, showing embodiments consistent with the present application, and together with the specification, are used to explain the principles of the present application. Obviously, the drawings described below are only some embodiments of the present application, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0042] Figure 1 A structural block diagram of an automated platform for lightweighting and performance evaluation of deep learning models according to an embodiment of the present application is shown;

[0043] Figure 2 A schematic diagram of a process of a model lightweighting method according to an embodiment of the present application is shown;

[0044] Figure 3 A schematic diagram of a flow chart of a model performance evaluation method according to an embodiment of the present application is shown;

[0045] Figure 4 A block diagram of an electronic device according to an embodiment of the present application is shown;

[0046] Figure 5 A schematic diagram of the structure of a computer system suitable for implementing an electronic device of an embodiment of the present application is shown. DETAILED DESCRIPTION

[0047] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this application will be more comprehensive and complete and fully convey the concept of the example embodiments to those skilled in the art.

[0048] In addition, described feature, structure or characteristic can be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present application. However, those skilled in the art will appreciate that the technical scheme of the present application can be put into practice without one or more of the specific details, or other methods, components, devices, steps, etc. can be adopted. In other cases, known methods, devices, realizations or operations are not shown or described in detail to avoid blurring the various aspects of the application.

[0049] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0050] The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps can be decomposed, and some operations / steps can be combined or partially combined, so the actual execution order may change according to actual conditions.

[0051] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only a part of the embodiment of the present invention, not all the embodiments. Based on the embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0052] Some embodiments of the present application will be described in detail below in conjunction with the accompanying drawings. In the absence of conflict, the following embodiments and features in the embodiments may be combined with each other.

[0053] See also Figure 1 , shows a structural block diagram of an automated platform for lightweighting and performance evaluation of deep learning models according to an embodiment of the present application.

[0054] like Figure 1 As shown, an automated platform for lightweight deep learning models and performance evaluation is presented, including: user interface module, algorithm management module, task scheduling module, model evaluation module and result analysis module;

[0055] The user interface module is respectively connected to the task scheduling module and the model evaluation module in communication, the task scheduling module is respectively connected to the algorithm management module and the model evaluation module in communication, and the model evaluation module is also connected to the result analysis module in communication;

[0056] The user interface module is used to upload models and data sets, configure lightweight tasks, and monitor task progress;

[0057] The algorithm management module is used to manage pre-integrated algorithms and user-defined algorithms, and provide algorithm packaging and deployment functions;

[0058] The task scheduling module is used to automatically schedule lightweight tasks and optimize resource allocation according to the task parameters configured by the user;

[0059] The model evaluation module is used to compare the effects of different lightweight solutions and provide performance reports;

[0060] The result analysis module is used to analyze lightweight results and provide a visual report.

[0061] It can be understood that the various modules in this embodiment communicate mainly through message queues and API interfaces to ensure efficient transmission and processing of data.

[0062] In some feasible embodiments, based on the aforementioned scheme, the algorithm management module includes an algorithm sub-module, and the algorithm sub-module has a built-in lightweight algorithm, and the lightweight algorithm includes a block pruning algorithm, a Qdrop quantization algorithm and a data-free knowledge distillation algorithm.

[0063] It should be noted that the algorithm submodule in this embodiment integrates common lightweight algorithms to ensure that users can choose appropriate algorithms according to their needs. In addition, in addition to the pre-integrated algorithms, users can also upload customized lightweight algorithms to the algorithm management module and encapsulate them as Docker containers to achieve personalized optimization.

[0064] In some feasible embodiments, based on the above-mentioned solution, the model evaluation module has a built-in model performance evaluation tool.

[0065] It should be noted that model performance evaluation tools refer to a series of methods and software used to measure the quality of models. Common tools include accuracy, recall and other indicator calculation tools. Accuracy measures the proportion of correct predictions made by the model, and recall reflects the model's ability to find all positive examples. There is also the mean square error (MSE), which is often used in regression models to calculate the average of the squared deviations between the predicted value and the true value. Cross-validation tools are also very important. They divide the data set into multiple parts, and use some of them for training and some for validation in turn, to more comprehensively evaluate the performance of the model on different data subsets. These tools can evaluate the accuracy, stability and other performance of the model from different angles, helping users improve the model or select the optimal model.

[0066] In summary, the automatic call platform provided in the embodiment of the present application has the following characteristics:

[0067] Multi-algorithm integration: The platform integrates common lightweight algorithms, such as block pruning algorithm, Qdrop quantization algorithm and data-free knowledge distillation algorithm. Users can choose the appropriate algorithm according to their needs.

[0068] User-defined algorithms: In addition to pre-integrated algorithms, users can also upload customized lightweight algorithms and encapsulate them as Docker containers to achieve personalized optimization.

[0069] Graphical User Interface (GUI): Provides an intuitive graphical interface that allows users to easily upload models and datasets, configure lightweight tasks, and monitor task progress.

[0070] Automated lightweight production line: A set of automated lightweight task production lines has been designed. Users can complete model lightweighting with one click through a preset configuration template, simplifying the operation process.

[0071] Model comparison and evaluation: The built-in model performance comparison and evaluation tool allows users to select different versions of models and test data sets, compare their performance through evaluation algorithms, and help users analyze the impact of different parameter settings on lightweight effects.

[0072] Simplified workflow: Through the graphical interface and automated production line, the technical difficulty of users participating in the lightweight process is reduced and work efficiency is improved.

[0073] Enhanced evaluation mechanism: The built-in model comparison and evaluation function allows users to easily compare the effects of different lightweight solutions and make the best choice.

[0074] Wide applicability: Both professional developers and non-professionals can use this platform to efficiently complete model optimization work, which is suitable for a variety of application scenarios.

[0075] Based on the same inventive concept, an embodiment of the present application further provides a model lightweight method, which is implemented based on an automated platform for lightweighting and performance evaluation of deep learning models described in any of the above embodiments. Specifically, the model lightweight method includes:

[0076] The user uploads the deep learning model file and corresponding dataset that need to be lightweight through the user interface module;

[0077] The user selects the required lightweight algorithm through the algorithm management module and configures the corresponding parameters according to the selected algorithm;

[0078] After configuring all parameters, start the lightweight task through the user interface module;

[0079] The task scheduling module receives the task request and assigns the task to the lightweight algorithm for optimization processing;

[0080] After processing, the optimized lightweight model is output.

[0081] In some feasible embodiments, based on the aforementioned solution, during the optimization process, the user monitors the task progress and views the task log through the user interface module.

[0082] For example, see Figure 2 , showing a flow chart of a model lightweighting method.

[0083] like Figure 2 As shown, the specific process of this method is as follows:

[0084] Step A1: User uploads model and dataset.

[0085] The user uploads the deep learning model files and corresponding data sets that need to be lightweight through the user interface module.

[0086] Step A2: Select a lightweight algorithm.

[0087] Users can choose a suitable lightweight algorithm on the platform. The platform provides a variety of pre-integrated lightweight algorithms, such as block pruning algorithm, Qdrop quantization algorithm and data-free knowledge distillation algorithm. Users can choose according to model characteristics and needs.

[0088] Step A3: Configure algorithm parameters.

[0089] Users configure the corresponding parameters according to the selected algorithm. The platform provides a parameter configuration wizard to help users understand the meaning and recommended values ​​of each parameter, ensuring the flexibility and personalization of the lightweight process.

[0090] Step A4: Start the lightweight task.

[0091] After configuring all parameters, the user starts the lightweight task through the GUI. The task scheduling module of the platform receives the task request and assigns the task to the corresponding lightweight algorithm for processing.

[0092] Step A5: Monitor task progress.

[0093] Users can view the progress of lightweight tasks in real time through the platform's monitoring interface. The platform provides detailed task status information, including processing speed, estimated completion time, etc., to ensure that users have a clear understanding of the lightweight process.

[0094] Step A6: Check the task log.

[0095] Users can view detailed logs of lightweight tasks to understand key events and potential problems during task execution. Log information helps users diagnose problems and verify results.

[0096] Step A7: The task is completed.

[0097] After the lightweight task is completed, the platform automatically stops the task and saves the lightweight model. Users can be notified of the task completion through the platform's notification system.

[0098] Step A8: Output the lightweight model.

[0099] The platform outputs the lightweight model to the storage location specified by the user. The lightweight model is optimized to have a smaller size and faster inference speed while maintaining a high accuracy.

[0100] Step A9: End.

[0101] After completing the lightweight process, users can download the lightweight model through the platform and perform further application deployment or testing.

[0102] Based on the same inventive concept, this embodiment further provides a model performance evaluation method, which is implemented based on an automated platform for lightweighting and performance evaluation of deep learning models described in any of the above embodiments. Specifically, the model performance evaluation method includes:

[0103] The user uploads the model and test data set that need to be compared and tested through the user interface module;

[0104] Users configure evaluation parameters and evaluation indicators according to evaluation needs;

[0105] After the configuration is completed, the evaluation task is started through the user interface module, and the task parameters are passed to the model evaluation module;

[0106] The model evaluation module evaluates the model according to the configured evaluation parameters and evaluation indicators and generates a detailed evaluation report.

[0107] For example, see Figure 3 , showing a flow chart of a model performance evaluation method.

[0108] like Figure 3 As shown, the specific process of this method is as follows:

[0109] Step B1: User selects model and dataset.

[0110] Users upload the models and test datasets that need to be compared and tested through the user interface module. The platform supports multiple model formats and dataset formats to ensure that users can easily choose.

[0111] Step B2: Configure evaluation parameters and evaluation indicators.

[0112] Users can configure evaluation parameters and evaluation indicators as needed. The platform provides predefined evaluation parameters and indicator templates, and users can also customize parameters and indicators.

[0113] Step B3: Start the evaluation task.

[0114] After completing the configuration, the user starts the evaluation task through the user interface module. The platform will automatically schedule the evaluation task and pass the task parameters to the model evaluation module.

[0115] Step B4: Generate an evaluation report.

[0116] The model evaluation module evaluates the model according to the parameters and indicators configured by the user and generates a detailed evaluation report. The report includes the model's performance indicators, comparative analysis, etc.

[0117] Step B5: Display the evaluation results.

[0118] The platform displays the evaluation results to users through the user interface module, and users can view detailed evaluation reports and comparative analysis results.

[0119] Step B6: End.

[0120] After the user completes the assessment, the user can end the assessment process through the user interface module. The platform will save the assessment results and reports for subsequent review and analysis by the user.

[0121] like Figure 4 As shown, an embodiment of the present application also provides an electronic device 400, including a memory 410, a processor 420, and a computer program 411 stored in the memory 410 and executable on the processor, and when the processor 420 executes the computer program 411, the steps of a model lightweight method or a model performance evaluation method mentioned above are implemented.

[0122] Figure 5 A schematic diagram of the structure of a computer system suitable for implementing an electronic device of an embodiment of the present application is shown.

[0123] It should be noted that Figure 5 The computer system 500 of the electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0124] like Figure 5 As shown, the computer system 500 includes a central processing unit (CPU) 501, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 502 or the program loaded from the storage part 508 to the random access memory (RAM) 503, such as executing the method described in the above embodiment. In the RAM 503, various programs and data required for system operation are also stored. The CPU 501, the ROM 502 and the RAM 503 are connected to each other through the bus 504. The input / output (I / O) interface 505 is also connected to the bus 504.

[0125] The following components are connected to the I / O interface 505: an input section 506 including a keyboard, a mouse, etc.; an output section 507 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker; a storage section 508 including a hard disk, etc.; and a communication section 509 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to the I / O interface 505 as needed. A removable medium 511, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 510 as needed so that a computer program read therefrom is installed into the storage section 508 as needed.

[0126] In particular, according to an embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication section 509, and / or installed from a removable medium 511. When the computer program is executed by a central processing unit (CPU) 501, various functions defined in the system of the present application are executed.

[0127] It should be noted that the computer-readable medium shown in the embodiment of the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, - but not limited to - an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by an instruction execution system, device or device or used in combination with it. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, wherein a computer-readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, which may send, propagate, or transmit programs for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0128] The flowchart and block diagram in the accompanying drawings illustrate the possible architecture, functions and operations of the system, method and computer program product according to various embodiments of the present application. Wherein, each box in the flowchart or block diagram can represent a module, a program segment, or a part of the code, and the above-mentioned module, program segment, or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0129] The units involved in the embodiments described in this application can be implemented in software or in hardware, and the described units can also be provided in a processor. Among them, the names of these units do not, in some cases, constitute a limitation on the units themselves.

[0130] As another aspect, this application also provides a computer program product or a computer program. The computer program product or the computer program includes computer instructions, and the computer instructions are stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes a model lightweighting method or a model performance evaluation method described in the above embodiments.

[0131] As another aspect, this application also provides a computer-readable medium. The computer-readable medium can be included in the electronic device described in the above embodiments; it can also exist separately without being assembled into the electronic device. The above computer-readable medium carries one or more programs. When the above one or more programs are executed by an electronic device, the electronic device implements a model lightweighting method or a model performance evaluation method described in the above embodiments.

[0132] It should be noted that although several modules or units of a device for action execution are mentioned in the above detailed description, such a division is not mandatory. In fact, according to the embodiments of this application, the features and functions of the two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0133] Through the description of the above embodiments, those skilled in the art can easily understand that the example embodiments described herein can be implemented by software or by a combination of software and necessary hardware. Therefore, the technical solutions according to the embodiments of this application can be embodied in the form of a software product. The software product can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, including several instructions to enable a computing device (which can be a personal computer, a server, a touch terminal, or a network device, etc.) to execute the method according to the embodiments of this application.

[0134] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the embodiments disclosed herein. The present application is intended to cover any variations, uses or adaptations of the present application, which follow the general principles of the present application and include common knowledge or customary techniques in the art that are not disclosed in the present application. It should be understood that the present application is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. An automated platform for lightweight deep learning models and performance evaluation, characterized in that: include: User interface module, algorithm management module, task scheduling module, model evaluation module and result analysis module; The user interface module is respectively connected to the task scheduling module and the model evaluation module in communication, the task scheduling module is respectively connected to the algorithm management module and the model evaluation module in communication, and the model evaluation module is also connected to the result analysis module in communication; The user interface module is used to upload models and data sets, configure lightweight tasks, and monitor task progress; The algorithm management module is used to manage pre-integrated algorithms and user-defined algorithms, and provide algorithm packaging and deployment functions; The task scheduling module is used to automatically schedule lightweight tasks and optimize resource allocation according to the task parameters configured by the user; The model evaluation module is used to compare the effects of different lightweight solutions and provide performance reports; The result analysis module is used to analyze lightweight results and provide a visual report.

2. The automation platform according to claim 1, characterized in that: The algorithm management module includes an algorithm submodule, and the algorithm submodule has a built-in lightweight algorithm. The lightweight algorithm includes a block pruning algorithm, a Qdrop quantization algorithm, and a data-free knowledge distillation algorithm.

3. The automation platform according to claim 1, characterized in that: The model evaluation module has a built-in model performance evaluation tool.

4. A model lightweight method implemented based on the automatic dialogue platform according to any one of claims 1 to 3, characterized in that: include: The user uploads the deep learning model file and corresponding dataset that need to be lightweight through the user interface module; The user selects the required lightweight algorithm through the algorithm management module and configures the corresponding parameters according to the selected algorithm; After configuring all parameters, start the lightweight task through the user interface module; The task scheduling module receives the task request and assigns the task to the lightweight algorithm for optimization processing; After processing, the optimized lightweight model is output.

5. The model lightweighting method according to claim 4, characterized in that: During the optimization process, users monitor the task progress and view the task log through the user interface module.

6. A computer-readable storage medium, characterized in that: The storage medium stores computer instructions, and when the computer instructions are executed on a computer, the computer executes the method according to any one of claims 4 to 5.

7. An electronic device, characterized in that: include: Memory and processor; The memory is used to store computer instructions; The processor is used to call the computer instructions stored in the memory so that the electronic device executes the method as described in any one of claims 4-5.

8. A model performance evaluation method implemented based on the automated speech platform according to any one of claims 1 to 3, characterized in that: include: The user uploads the model and test data set that need to be compared and tested through the user interface module; Users configure evaluation parameters and evaluation indicators according to evaluation needs; After the configuration is completed, the evaluation task is started through the user interface module, and the task parameters are passed to the model evaluation module; The model evaluation module evaluates the model according to the configured evaluation parameters and evaluation indicators and generates a detailed evaluation report.

9. A computer-readable storage medium, characterized in that: The storage medium stores computer instructions, which, when executed on a computer, enable the computer to execute the method according to claim 8.

10. An electronic device, characterized in that: include: Memory and processor; The memory is used to store computer instructions; The processor is used to call the computer instructions stored in the memory so that the electronic device executes the method as claimed in claim 8.

Images