Visualization system, method and device for model training process
By introducing a visual system during model training, the learning curve is displayed in real time, which solves the problem that users find it difficult to understand the training state at all times, and achieves faster strategy adjustment and reduces training costs.
Patent Information
- Application Number
- CN202411717521.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-27
- Publication Date
- 2025-04-25
AI Technical Summary
Existing model training techniques cannot display training status at any time, making it difficult for users to quickly adjust strategies during the training process, such as adjusting the number of iterations or model structure, which may lead to overfitting and increasing training costs.
Design a visualization system for model training process, including data selection module, model training module and training monitoring module, and realize the visualization of modeling and training process by displaying the learning curve of the model in real time.
By displaying the learning curve at any time, users can promptly understand the training situation, assist in adjusting the model topology and training iterations, reduce training time and cost, and avoid overfitting.
Smart Images
Figure CN119883453A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of model training technology, and specifically to a visualization system for a model training process, a visualization method for a model training process, a visualization device for a model training process, a machine-readable storage medium, and an electronic device. Background Art
[0002] The existing model training is mainly completed by developers in the background. Users cannot obtain the real-time status of the model training process. Usually, the parameter values related to the model can only be obtained after the training is completed. This is not conducive to quickly adjusting the model strategy during the training process. For example, after the training model stabilizes and reaches a certain accuracy, stop training, change the model's neural network structure; or adjust the number of training iterations. Because the prediction accuracy of the model does not increase with the increase in the number of iterations, on the contrary, the high model training accuracy caused by too high a number of iterations will cause the model to overfit, which will increase the cost of model training while reducing the generalization ability of the model. Therefore, during the model training process, users cannot always know the training status. Summary of the invention
[0003] The purpose of the embodiments of the present invention is to provide a visualization system for a model training process, a visualization method for a model training process, a visualization device for a model training process, a machine-readable storage medium and an electronic device. The visualization system for the model training process realizes the visualization of the modeling and training process by displaying the learning curve of the model at all times, so that the user can know the training status at all times.
[0004] In order to achieve the above-mentioned object, the first aspect of the present application provides a visualization system for a model training process, including: a data selection module, a model training module and a training monitoring module;
[0005] The data selection module is used to obtain parameters required for model training and call the model training module in response to a model training instruction;
[0006] The model training module is used to train the preset initial model according to the parameters required for the model training, obtain real-time training data, and send the real-time training data to the training monitoring module;
[0007] The training monitoring module is used to draw a learning curve based on the real-time training data, and display the learning curve in real time.
[0008] In an embodiment of the present application, the model training module is used to train a preset initial model in the first thread based on the parameters required for the model training, obtain real-time training data, and send the real-time training data to the second thread through a semaphore;
[0009] The training monitoring module is used to draw a learning curve based on the real-time training data in the second thread, and display the learning curve in real time.
[0010] In an embodiment of the present application, the real-time training data includes at least one indicator data;
[0011] The training monitoring module is used to draw corresponding learning curves based on various indicator data, and to display the learning curves corresponding to the various indicator data in real time.
[0012] In an embodiment of the present application, a data preview module is also included, and the data preview module is used to display the parameters required for the model training.
[0013] In the embodiment of the present application, a message output module is also included;
[0014] The model training module is also used to generate a model training log and send the model training log to the message output module for display.
[0015] A second aspect of the present application provides a method for visualizing a model training process, comprising:
[0016] Get the parameters required for model training;
[0017] In response to the model training instruction, the preset initial model is trained according to the parameters required for the model training to obtain real-time training data;
[0018] Based on the real-time training data, a learning curve is drawn, and the learning curve is displayed in real time.
[0019] In the embodiment of the present application, the preset initial model is trained according to the parameters required for the model training to obtain real-time training data, including:
[0020] In the first thread, based on the parameters required for the model training, the preset initial model is trained to obtain real-time training data, and the real-time training data is sent to the second thread through a semaphore;
[0021] The step of drawing a learning curve based on the real-time training data and displaying the learning curve in real time includes:
[0022] In the second thread, a learning curve is drawn based on the real-time training data, and the learning curve is displayed in real time.
[0023] In an embodiment of the present application, the real-time training data includes at least one indicator data;
[0024] The step of drawing a learning curve based on the real-time training data and displaying the learning curve in real time includes:
[0025] Corresponding learning curves are drawn based on various indicator data respectively, and the learning curves corresponding to the various indicator data are displayed in real time respectively.
[0026] In the embodiment of the present application, the step of obtaining the parameters required for model training includes:
[0027] Acquire initial parameters, which include sample data and model parameters required for model training;
[0028] The initial parameters are preprocessed to obtain parameters required for model training.
[0029] In the embodiment of the present application, it also includes:
[0030] In response to the model export instruction, the current model is exported using the Verilog-A language to obtain a final model.
[0031] A third aspect of the present application is a visualization device for a model training process, comprising:
[0032] Acquisition module, used to obtain the parameters required for model training;
[0033] A training module, for responding to a model training instruction, training a preset initial model according to parameters required for the model training, and obtaining real-time training data;
[0034] A drawing module is used to draw a learning curve based on the real-time training data, and display the learning curve in real time.
[0035] In an embodiment of the present application, the training module includes a first thread unit;
[0036] The first thread unit is used to train the preset initial model in the first thread based on the parameters required for the model training to obtain real-time training data, and send the real-time training data to the second thread through a semaphore;
[0037] The drawing module includes a second thread unit;
[0038] The second thread unit is used to draw a learning curve based on the real-time training data in the second thread, and display the learning curve in real time.
[0039] A fourth aspect of the present application provides an electronic device, the electronic device comprising:
[0040] at least one processor;
[0041] a memory connected to the at least one processor;
[0042] The memory stores instructions that can be executed by the at least one processor, and the at least one processor implements the above-mentioned visualization method of the model training process by executing the instructions stored in the memory.
[0043] A fifth aspect of the present application provides a machine-readable storage medium having instructions stored thereon, which, when executed by a processor, configure the processor to execute the above-mentioned visualization method of the model training process.
[0044] Through the above technical solution, by setting up a data selection module, a model training module and a training monitoring module; the data selection module obtains the parameters required for model training, and calls the model training module in response to the model training instruction; the model training module trains the preset initial model according to the parameters required for model training, obtains real-time training data, and sends the real-time training data to the training monitoring module; the training monitoring module draws a learning curve based on the real-time training data, and displays the learning curve in real time. By displaying the learning curve of the model from time to time, the visualization of the modeling and training process is realized, so that the user can know the training status from time to time, and then according to the changing trend of the learning curve, the user can be assisted to make model topology structure adjustment and training iteration number optimization strategy, thereby reducing training time and cost.
[0045] Other features and advantages of the embodiments of the present invention will be described in detail in the subsequent detailed description. BRIEF DESCRIPTION OF THE DRAWINGS
[0046] The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following specific embodiments, they are used to explain the embodiments of the present invention, but do not constitute a limitation on the embodiments of the present invention. In the accompanying drawings:
[0047] Figure 1 A schematic diagram of a visualization system for a model training process according to an embodiment of the present application is schematically shown;
[0048] Figure 2 A schematic diagram of feature data X and label data Y required for setting a model for a user GUI interface according to an embodiment of the present application is shown;
[0049] Figure 3 A schematic diagram of setting the model data set ratio, number of iterations and neural network structure according to a user GUI interface of an embodiment of the present application is shown;
[0050] Figure 4A schematic diagram of a learning curve in a training process of a display model according to an embodiment of the present application is schematically shown;
[0051] Figure 5 A schematic diagram of a flow chart of a visualization method for a model training process according to an embodiment of the present application is schematically shown;
[0052] Figure 6 A schematic diagram of the structure of a visualization device for a model training process according to an embodiment of the present application is shown;
[0053] Figure 7 The internal structure diagram of a computer device according to an embodiment of the present application is schematically shown.
[0054] Description of Reference Numerals
[0055] 410 - acquisition module; 420 - training module; 430 - drawing module; A01 - processor; A02 - network interface; A03 - internal memory; A04 - display screen; A05 - input device; A06 - non-volatile storage medium; B01 - operating system; B02 - computer program. DETAILED DESCRIPTION
[0056] The specific implementation of the embodiment of the present invention is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described here is only used to illustrate and explain the embodiment of the present invention, and is not used to limit the embodiment of the present invention.
[0057] It should be noted that the acquisition, transmission, storage, use, and processing of data in the technical solution of this application are in compliance with the relevant provisions of national laws and regulations. In the embodiments of this application, some existing solutions in the industry such as certain software, components, and models may be mentioned, which should be considered as exemplary. Their purpose is only to illustrate the feasibility of implementing the technical solution of this application, but it does not mean that the applicant has or will necessarily use the solution.
[0058] It should be noted that if the embodiments of the present application involve directional indications (such as up, down, left, right, front, back...), the directional indications are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.
[0059] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present application, the descriptions of "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or suggesting their relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by this application.
[0060] The abbreviations and key terms mentioned in this embodiment are defined as follows:
[0061] PyQT: is a Python binding Digia QT application framework that provides rich functions and powerful graphical user interface (GUI) development capabilities;
[0062] QDockWidget: is a class in the Qt framework;
[0063] Matplotlib: is a powerful data visualization library for Python.
[0064] GUI: Graphical User Interface;
[0065] TCAD: Technology Computer Aided Design;
[0066] SPICE: SPICE (Simulation Program with Integrated Circuit Emphasis) is a powerful general-purpose circuit-level analog simulator;
[0067] Scikit-Learn: is a Python library for machine learning.
[0068] PyTorch: PyTorch is an open source machine learning library widely used in the field of deep learning.
[0069] Seaborn: is a Python-based data visualization library built on Matplotlib, providing a more advanced interface and more beautiful graphic style;
[0070] Qthread: In PyQt, QThread is a class used to implement multithreaded programming;
[0071] Python threading: In Python, the threading module is used to implement multithreaded programming;
[0072] pyqtSignal: In PyQt, pyqtSignal is a special type used to implement the signal and slot mechanism.
[0073] Please refer to Figure 1 , Figure 1 A schematic diagram of a visualization system for a model training process according to an embodiment of the present application is schematically shown. This embodiment provides a visualization system for a model training process, including: a data selection module, a model training module and a training monitoring module;
[0074] The data selection module is used to obtain parameters required for model training and call the model training module in response to a model training instruction;
[0075] The model training module is used to train the preset initial model according to the parameters required for the model training, obtain real-time training data, and send the real-time training data to the training monitoring module;
[0076] The training monitoring module is used to draw a learning curve based on the real-time training data, and display the learning curve in real time.
[0077] In this embodiment, the multiple document interface (MDI) technology of PyQT can be utilized to use QDockWidget to instantiate the underlying view modules of the visualization system including the console, that is, to generate a data selection module and a model training module, and at the same time integrate Matplotlib to form a view area specifically for graphics drawing, that is, to generate a training monitoring module.
[0078] The data selection module can be a module that allows users to select data items to obtain the parameters required for model training, including feature data, label data, number of iterations, etc. For example, please refer to Figure 2 and Figure 3 , the user GUI interface sets the feature data X and label data Y required for the model, and the user GUI interface sets the segmentation ratio, number of iterations, training method, device name, number of hidden layers, etc. of the model data set. The model training instruction can be automatically generated after the acquisition of the parameters required for model training is completed, or a model training button can be set in the data selection module. After the user enters the parameters required for model training, the user can press the training button, and then respond to the model training instruction to call the model training module to start the training model.
[0079] Take the training of circuit simulation prediction model as an example: input device TCAD simulation data and circuit SPICE data through the data selection module, and use the data selection module to let the user specify the feature data X and label data Y required for model training on the GUI interface, as well as the parameters required for training, including learning rate, network model architecture, etc. and the number of model training iterations (epochs), to train the prediction model.
[0080] It should be noted that the data selection module can also use Scikit-Learn and other technologies to pre-process the parameters required for model training, including standardization, regularization, and normalization. Regularization includes L1 regularization and L2 regularization.
[0081] The model training module starts training the model according to the parameters required for the model training. Specifically, it can be to use machine learning and artificial intelligence technologies such as PyTorch and Scikit-learn to carry out model training. The initial model can be a model established in advance as needed, such as: establishing a machine learning or neural network prediction model. During the training process, the data obtained from each training is used as real-time training data, and the real-time training data is sent to the training monitoring module. It should be noted that model training can be achieved using existing technologies, which will not be repeated here.
[0082] After obtaining the real-time training data, the training monitoring module executes the drawing function and passes its result to a view area dedicated to graphics drawing based on Matplotlib, Seaborn, etc. as a tool for drawing model training curves for real-time display of the learning curve.
[0083] In the above implementation process, a data selection module, a model training module and a training monitoring module are set; the data selection module obtains the parameters required for model training, and calls the model training module in response to the model training instruction; the model training module trains the preset initial model according to the parameters required for model training, obtains real-time training data, and sends the real-time training data to the training monitoring module; the training monitoring module draws a learning curve based on the real-time training data, and displays the learning curve in real time. By displaying the learning curve of the model from time to time, the visualization of the modeling and training process is achieved, so that the user can always know the training status, and then according to the changing trend of the learning curve, the user can be assisted in making model topology structure adjustments and training iteration number optimization strategies, thereby reducing training time and cost.
[0084] In some embodiments, the model training module is used to train a preset initial model in the first thread based on the parameters required for the model training to obtain real-time training data, and send the real-time training data to the second thread through a semaphore;
[0085] The training monitoring module is used to draw a learning curve based on the real-time training data in the second thread, and display the learning curve in real time.
[0086] In this embodiment, the model training and learning curve drawing can be placed in different threads, and semaphores can be used for inter-thread communication; that is, the model training module is executed in the first thread, and the training monitoring module is executed in the second thread, and the two do not interfere with each other. The multi-threading technology can be QT's own Qthread, Python's threading, etc., and the inter-thread semaphore communication can use pyqtSignal, and the real-time training data required for drawing is passed to the drawing function through the semaphore to draw the learning curve.
[0087] By placing the model training and learning curve drawing in different threads, the two do not interfere with each other, improving the visualization performance and efficiency and further reducing the training time.
[0088] In some embodiments, the real-time training data includes at least one indicator data;
[0089] The training monitoring module is used to draw corresponding learning curves based on various indicator data, and to display the learning curves corresponding to the various indicator data in real time.
[0090] In this embodiment, the above-mentioned indicator data include mean absolute error (MAE), mean square error (MSE), root mean square error (RMSE), and determination coefficient (R-Square, R2). Among them:
[0091]
[0092] Among them, m is the number of iterations, i is the current training number, and y i is the model output. Figure 4 , Figure 4 The figure schematically shows a learning curve diagram of the real-time display model training process according to an embodiment of the present application. The figure shows the learning curve of the mean absolute error MAE and the coefficient of determination R2.
[0093] The training monitoring module draws corresponding learning curves based on various indicator data, and displays the learning curves corresponding to the various indicator data in real time. The model can be evaluated according to different indicators, so as to have a more comprehensive understanding of the model training situation.
[0094] In some embodiments, a data preview module is also included, and the data preview module is used to display the parameters required for the model training.
[0095] In this embodiment, the data preview module can display the parameters required for the input model training in a list format for easy viewing by the user. It should be noted that the data preview module can also pre-process the parameters required for model training using technologies such as Scikit-Learn, including standardization, regularization, and normalization. Regularization includes L1 regularization and L2 regularization.
[0096] In some embodiments, a message output module is also included;
[0097] The model training module is also used to generate a model training log and send the model training log to the message output module for display.
[0098] In this embodiment, the message output module can display the model training log so that the user can view the detailed training status.
[0099] Please see Figure 5 , Figure 5 The flowchart of a visualization method of a model training process according to an embodiment of the present application is schematically shown. This embodiment provides a visualization method of a model training process, comprising the following steps:
[0100] Step 210: Obtain parameters required for model training;
[0101] In this embodiment, the above acquisition can be achieved by the user selecting data items to obtain parameters required for model training, including feature data, label data, number of iterations, etc.
[0102] In some embodiments, obtaining parameters required for model training includes:
[0103] First, initial parameters are obtained, wherein the initial parameters include sample data and model parameters required for model training;
[0104] Then, the initial parameters are preprocessed to obtain the parameters required for model training.
[0105] In this embodiment, the initial parameters can be preprocessed using Scikit-Learn and other technologies, including standardization, regularization, and normalization. Regularization includes L1 regularization and L2 regularization. By preprocessing the initial parameters, the parameters required for model training are more consistent with model training, which helps to improve the efficiency of model training.
[0106] Step 220: In response to the model training instruction, the preset initial model is trained according to the parameters required for the model training to obtain real-time training data;
[0107] In this embodiment, the above-mentioned model training instruction can be automatically generated after the acquisition of the parameters required for model training is completed, or a model training button can be set. After the user enters the parameters required for model training, the model training button is pressed to respond to the model training instruction. Model training can be carried out using machine learning and artificial intelligence technologies such as PyTorch and Scikit-learn. The initial model can be a model established in advance as needed, such as: establishing a machine learning or neural network prediction model. During the training process, the data obtained from each training is used as real-time training data. It should be noted that model training can be achieved using existing technologies, which will not be repeated here.
[0108] Step 230: Draw a learning curve based on the real-time training data, and display the learning curve in real time.
[0109] In this embodiment, after obtaining the real-time training data, the drawing function is executed and its result is passed to a view area specially used for graphics drawing formed based on Matplotlib, Seaborn, etc. as a tool for drawing model training curves to display the learning curve in real time.
[0110] In some embodiments, the real-time training data includes at least one indicator data; accordingly, drawing a learning curve based on the real-time training data and displaying the learning curve in real time includes: drawing corresponding learning curves based on various indicator data, and displaying the learning curves corresponding to the various indicator data in real time.
[0111] In this embodiment, the above-mentioned indicator data include mean absolute error (MAE), mean square error (MSE), root mean square error (RMSE), and determination coefficient (R-Square, R2). By drawing corresponding learning curves based on various indicator data, and displaying the learning curves corresponding to the various indicator data in real time, the model can be evaluated according to different indicators, so as to understand the model training situation more comprehensively.
[0112] In the above implementation process, the parameters required for model training are obtained; in response to the model training instruction, the preset initial model is trained according to the parameters required for model training to obtain real-time training data; based on the real-time training data, a learning curve is drawn and the learning curve is displayed in real time. By displaying the learning curve of the model at all times, the visualization of the modeling and training process is achieved, so that the user can know the training status at all times, and then according to the changing trend of the learning curve, the user can be assisted in making model topology structure adjustments and training iteration optimization strategies, thereby reducing training time and cost.
[0113] In some embodiments, the step of training a preset initial model based on the parameters required for model training to obtain real-time training data includes: in a first thread, training a preset initial model based on the parameters required for model training to obtain real-time training data, and sending the real-time training data to a second thread via a semaphore;
[0114] Correspondingly, drawing a learning curve based on the real-time training data and displaying the learning curve in real time includes: in the second thread, drawing a learning curve based on the real-time training data and displaying the learning curve in real time.
[0115] In this embodiment, the model training and learning curve drawing can be placed in different threads, and semaphores can be used for inter-thread communication; that is, the model training module is executed in the first thread, and the training monitoring module is executed in the second thread, and the two do not interfere with each other. The multi-threading technology can be QT's own Qthread, Python's threading, etc., and the inter-thread semaphore communication can use pyqtSignal, and the real-time training data required for drawing is passed to the drawing function through the semaphore to draw the learning curve.
[0116] By placing the model training and learning curve drawing in different threads, the two do not interfere with each other, improving the visualization performance and efficiency and further reducing the training time.
[0117] In some embodiments, the method further includes: in response to a model export instruction, exporting the current model using Verilog-A language to obtain a final model.
[0118] In this embodiment, Verilog-A is a widely accepted hardware description language, so that the final model can be used on different tool platforms, thereby improving the portability of the design.
[0119] Figure 5 FIG. 1 is a flow chart of a visualization method for the model training process in an embodiment. It should be understood that although Figure 5The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 5 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.
[0120] Please see Figure 6 , Figure 6 The structure diagram of a visualization device for a model training process according to an embodiment of the present application is schematically shown. This embodiment provides a visualization device for a model training process, including an acquisition module 410, a training module 420 and a drawing module 430, wherein:
[0121] An acquisition module 410 is used to acquire parameters required for model training;
[0122] The training module 420 is used to respond to the model training instruction and train the preset initial model according to the parameters required for the model training to obtain real-time training data;
[0123] The drawing module 430 is used to draw a learning curve based on the real-time training data, and display the learning curve in real time.
[0124] Wherein, the training module includes a first thread unit;
[0125] The first thread unit is used to train the preset initial model in the first thread based on the parameters required for the model training to obtain real-time training data, and send the real-time training data to the second thread through a semaphore;
[0126] The drawing module includes a second thread unit;
[0127] The second thread unit is used to draw a learning curve based on the real-time training data in the second thread, and display the learning curve in real time.
[0128] The visualization device of the model training process includes a processor and a memory. The above-mentioned acquisition module 410, training module 420 and drawing module 430 are all stored in the memory as program units, and the processor executes the above-mentioned program units stored in the memory to realize corresponding functions.
[0129] The processor contains a kernel, which calls the corresponding program unit from the memory. One or more kernels can be set, and the visualization of the modeling and training process can be achieved by adjusting the kernel parameters.
[0130] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.
[0131] An embodiment of the present invention provides a machine-readable storage medium having a program stored thereon, which, when executed by a processor, implements the visualization method of the model training process.
[0132] An embodiment of the present invention provides a processor, which is used to run a program, wherein the program executes a visualization method of the model training process when running.
[0133] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 7 As shown. The computer device includes a processor A01, a network interface A02, a display screen A04, an input device A05 and a memory (not shown in the figure) connected via a system bus. Among them, the processor A01 of the computer device is used to provide computing and control capabilities. The memory of the computer device includes an internal memory A03 and a non-volatile storage medium A06. The non-volatile storage medium A06 stores an operating system B01 and a computer program B02. The internal memory A03 provides an environment for the operation of the operating system B01 and the computer program B02 in the non-volatile storage medium A06. The network interface A02 of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor A01, a method for visualizing a model training process is implemented. The display screen A04 of the computer device can be a liquid crystal display or an electronic ink display, and the input device A05 of the computer device can be a touch layer covered on the display screen, or a button, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.
[0134] Those skilled in the art will understand that Figure 7 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.
[0135] In one embodiment, the visualization method and apparatus of the model training process provided in the present application can be implemented in the form of a computer program, which can be used in Figure 7 The computer device shown in the figure is run. The memory of the computer device can store various program modules constituting the visualization method device of the model training process, such as: Figure 6 The acquisition module 410, training module 420 and drawing module 430 are shown. The computer program composed of various program modules enables the processor to execute the steps of the visualization method of the model training process of each embodiment of the present application described in this specification.
[0136] Figure 7 The computer device shown can be Figure 6 The acquisition module 410 in the visualization method device of the model training process shown executes step 210. The computer device may execute step 220 through the training module 420. The computer device may execute step 230 through the drawing module 430.
[0137] An embodiment of the present application provides an electronic device, the electronic device comprising: at least one processor; a memory connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the at least one processor implements the above-mentioned visualization method of the model training process by executing the instructions stored in the memory, and the processor implements the following steps when executing the instructions:
[0138] Get the parameters required for model training;
[0139] In response to the model training instruction, the preset initial model is trained according to the parameters required for the model training to obtain real-time training data;
[0140] Based on the real-time training data, a learning curve is drawn, and the learning curve is displayed in real time.
[0141] In one embodiment, the training of the preset initial model according to the parameters required for the model training to obtain real-time training data includes:
[0142] In the first thread, based on the parameters required for the model training, the preset initial model is trained to obtain real-time training data, and the real-time training data is sent to the second thread through a semaphore;
[0143] The step of drawing a learning curve based on the real-time training data and displaying the learning curve in real time includes:
[0144] In the second thread, a learning curve is drawn based on the real-time training data, and the learning curve is displayed in real time.
[0145] In one embodiment, the real-time training data includes at least one indicator data;
[0146] The step of drawing a learning curve based on the real-time training data and displaying the learning curve in real time includes:
[0147] Corresponding learning curves are drawn based on various indicator data respectively, and the learning curves corresponding to the various indicator data are displayed in real time respectively.
[0148] In one embodiment, obtaining parameters required for model training includes:
[0149] Acquire initial parameters, which include sample data and model parameters required for model training;
[0150] The initial parameters are preprocessed to obtain parameters required for model training.
[0151] In one embodiment, it further includes:
[0152] In response to the model export instruction, the current model is exported using the Verilog-A language to obtain a final model.
[0153] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.
[0154] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0155] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0156] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.
[0157] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.
[0158] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.
[0159] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.
[0160] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.
[0161] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.
Claims
1. A visualization system for a model training process, characterized in that: include: Data selection module, model training module and training monitoring module; The data selection module is used to obtain parameters required for model training and call the model training module in response to a model training instruction; The model training module is used to train the preset initial model according to the parameters required for the model training, obtain real-time training data, and send the real-time training data to the training monitoring module; The training monitoring module is used to draw a learning curve based on the real-time training data, and display the learning curve in real time.
2. The visualization system for model training process according to claim 1, characterized in that: The model training module is used to train a preset initial model in the first thread based on the parameters required for the model training to obtain real-time training data, and send the real-time training data to the second thread through a semaphore; The training monitoring module is used to draw a learning curve based on the real-time training data in the second thread, and display the learning curve in real time.
3. The visualization system for model training process according to claim 1, characterized in that: The real-time training data includes at least one indicator data; The training monitoring module is used to draw corresponding learning curves based on various indicator data, and to display the learning curves corresponding to the various indicator data in real time.
4. The visualization system for model training process according to claim 1, characterized in that: It also includes a data preview module, which is used to display the parameters required for the model training.
5. The visualization system for model training process according to claim 1, characterized in that: Also includes a message output module; The model training module is also used to generate a model training log and send the model training log to the message output module for display.
6. A method for visualizing a model training process, characterized in that: include: Get the parameters required for model training; In response to the model training instruction, the preset initial model is trained according to the parameters required for the model training to obtain real-time training data; Based on the real-time training data, a learning curve is drawn, and the learning curve is displayed in real time.
7. The method for visualizing a model training process according to claim 6, characterized in that: The step of training a preset initial model according to the parameters required for the model training to obtain real-time training data includes: In the first thread, based on the parameters required for the model training, the preset initial model is trained to obtain real-time training data, and the real-time training data is sent to the second thread through a semaphore; The step of drawing a learning curve based on the real-time training data and displaying the learning curve in real time includes: In the second thread, a learning curve is drawn based on the real-time training data, and the learning curve is displayed in real time.
8. The method for visualizing a model training process according to claim 6, characterized in that: The real-time training data includes at least one indicator data; The step of drawing a learning curve based on the real-time training data and displaying the learning curve in real time includes: Corresponding learning curves are drawn based on various indicator data respectively, and the learning curves corresponding to the various indicator data are displayed in real time respectively.
9. The method for visualizing a model training process according to claim 6, characterized in that: The obtaining of parameters required for model training includes: Acquire initial parameters, which include sample data and model parameters required for model training; The initial parameters are preprocessed to obtain parameters required for model training.
10. The method for visualizing a model training process according to claim 6, characterized in that: Also includes: In response to the model export instruction, the current model is exported using the Verilog-A language to obtain a final model.
11. A visualization device for a model training process, characterized in that: include: Acquisition module, used to obtain the parameters required for model training; A training module, for responding to a model training instruction, training a preset initial model according to parameters required for the model training, and obtaining real-time training data; A drawing module is used to draw a learning curve based on the real-time training data, and display the learning curve in real time.
12. The visualization device for model training process according to claim 11, characterized in that: The training module includes a first thread unit; The first thread unit is used to train the preset initial model in the first thread based on the parameters required for the model training to obtain real-time training data, and send the real-time training data to the second thread through a semaphore; The drawing module includes a second thread unit; The second thread unit is used to draw a learning curve based on the real-time training data in the second thread, and display the learning curve in real time.
13. An electronic device, characterized in that: The electronic device includes: at least one processor; a memory connected to the at least one processor; The memory stores instructions that can be executed by the at least one processor, and the at least one processor implements the visualization method of the model training process described in any one of claims 6 to 10 by executing the instructions stored in the memory.
14. A machine-readable storage medium having instructions stored thereon, characterized in that: When the instruction is executed by a processor, the processor is configured to perform a visualization method for a model training process according to any one of claims 6 to 10.