A large model-based running training method, device and storage medium

Abstract

The application relates to a large model-based running training method and device, and belongs to the technical field of artificial intelligence. The method comprises the following steps: pre-training a preset first large model, recording multi-modal data of an athlete, inputting the multi-modal data into the pre-trained large model, and obtaining a running training scheme of the athlete. The application also provides a large model-based running training device and a storage medium.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of artificial intelligence, and particularly relates to a running training method and device based on a large model and a storage medium. BACKGROUND

[0002] With the increasing emphasis on health education, the analysis of sports training data is becoming increasingly important. Current data applications mainly focus on data collection and basic physical fitness testing, and the guidance for after-school training is relatively lacking. Usually, it is not possible to dynamically adjust the training according to the individual physical fitness, training state and training effect of students. This leads to homogeneity and inefficiency of training guidance, and cannot fully meet the needs of individualized training of students. In the training of professional athletes, it is also necessary to develop individualized training methods according to the individual physical fitness of athletes and their past training performance, so as to provide more scientific training programs and improve sports performance.

[0003] With the development of artificial intelligence, large models have strong capabilities in natural language processing, image analysis and multi-modal data fusion. How to use large models to provide individualized and real-time training programs based on multi-modal data such as athlete's physical fitness data, training images and training evaluation results is a technical problem that needs to be solved at present. SUMMARY

[0004] In view of the deficiencies of the prior art, the purpose of the application is to provide a running training method and device based on a large model, which improves the efficiency of running training, provides an individualized and real-time running training method, and realizes scientific and efficient physical fitness improvement.

[0005] In a first aspect of the application, a running training method based on a large model is provided, comprising:

[0006] S1, pre-training a preset first large model;

[0007] S2, recording multi-modal data of an athlete;

[0008] S3, inputting the multi-modal data into the pre-trained first large model to obtain a running training program for the athlete;

[0009] The pre-training of the preset first large model comprises:

[0010] Recording historical multi-modal data of the athlete's running;

[0011] Recording a training program corresponding to each piece of data in the historical multi-modal data;

[0012] Importing the historical multi-modal data and the training program into the preset first large model to construct a fine-tuning data set of the running training program;

[0013] wherein the multi-modal data comprises one or a combination of the following: running ability evaluation result, height, weight, latest running performance score, gender, age.

[0014] Further, the running ability evaluation result M1 is determined according to the following expression:

[0015] M1 = "Running ability evaluation:" + g1 + g2 + g3 + g4 + g5;

[0016] wherein g1 is a posture standard evaluation result;

[0017] g2 is an explosive power evaluation result;

[0018] g3 is an endurance evaluation result;

[0019] g4 is a running skill evaluation result;

[0020] g5 is a training quality evaluation result;

[0021] wherein the running ability evaluation result M1 is string data, the + in the expression represents the concatenation of strings, and g1, g2, g3, g4 and g5 are string data.

[0022] Further, the posture standard evaluation result g1 is determined according to the following method:

[0023] pre-training a preset second large model using historical running posture data of the athlete;

[0024] record the running posture images of the athlete for consecutive N1 days, identify each running posture image through the pre-trained second large model, output whether the posture of the runner conforms to the standard, and record the number n1 of images conforming to the standard and the number n2 of images not conforming to the standard;

[0025] determine the posture standard evaluation result g1 according to the following expression:

[0026]

[0027] wherein ts1 is a set first judgment threshold, ts2 is a set second judgment threshold, ts3 is a set third judgment threshold, and N1 is an integer greater than or equal to 20.

[0028] Further, the explosive power evaluation result g2 is determined according to the following method:

[0029] record the short-distance running item scores g2 of the athlete for consecutive N2 days i wherein i is the serial number of the short-distance running item score, i = 1, 2, 3,..., n3, and n3 is the number of short-distance running item scores of the athlete for consecutive N2 days.

[0030] The explosive power evaluation result g2 is determined according to the following expression:

[0031]

[0032] wherein g2 is the explosive power evaluation score, and g2 is determined by the following expression: 21 21

[0033]

[0034] wherein g21 is the sub-zone score, and g21 is determined by the following expression: i i

[0035]

[0036] wherein N2 is an integer greater than or equal to 20.

[0037] Further, the endurance evaluation result g3 is determined according to the following method:

[0038] The long-distance running item scores g3 of the athlete for consecutive N3 days are recorded j wherein j is the serial number of the long-distance running item score, j = 1, 2, 3,..., n4, and n4 is the number of long-distance running item scores of the athlete for consecutive N3 days;

[0039] The endurance evaluation result g3 is determined according to the following expression:

[0040]

[0041] wherein g3 is the endurance evaluation score and is determined by the following expression: 31

[0042]

[0043] wherein N3 is an integer greater than or equal to 20.

[0044] Further, the running skill evaluation result g4 is determined according to the following method:

[0045] The time t of the athlete to complete each lap of the long-distance running item within N4 days is recorded m , the corresponding total number of laps n5, the distance of each lap l1, m = 1,..., n5, and the serial number of the completed laps;

[0046]

[0047] g4 = pace stability, good skill;

[0048] when​​​​​ and and when:

[0049] g4= pace control is very unstable ups and downs, lack of running skills;

[0050] when:

[0051] g4= pace fluctuation, running skills need to be further strengthened;

[0052] wherein ts4 is a set fourth judgment threshold, ts5 is a set fifth judgment threshold, ts6 is a set sixth judgment threshold;

[0053] wherein N4 is an integer greater than or equal to 20; p=1,...,n5, is the serial number of the number of laps completed.

[0054] Further, the training quality evaluation result g5 is determined according to the following method:

[0055] Record the running results g5 of the athlete for consecutive N5 days p , p=1,...,n3+n4…, is the serial number of the running, the smaller the serial number, the earlier the training record time, n3 is the number of short-distance running item scores of the athlete for consecutive N2 days, n4 is the number of long-distance running item scores of the athlete for consecutive N3 days, wherein N2 is an integer greater than or equal to 20, N3 is an integer greater than or equal to 20;

[0056] The training quality evaluation result g5 is determined according to the following expression:

[0057] g5=g 51 +g 52 +g 53 , wherein the plus sign represents string concatenation;

[0058] wherein g 51 is the quantity evaluation, g 52 is the quality evaluation, g 53 is the effect evaluation, and satisfies the following expression:

[0059]

[0060] wherein ts7 is a set seventh judgment threshold, ts8 is a set eighth judgment threshold, ts9 is a set ninth judgment threshold, ts 10 is a set tenth judgment threshold, ts 11 is a set eleventh judgment threshold, ts 12 is a set twelfth judgment threshold, ts 13 is a set thirteenth judgment threshold, ts 14 is a set fourteenth judgment threshold, ts15 a fifteenth judgment threshold value ts 16 a sixteenth judgment threshold value ts 17 a seventeenth judgment threshold value ts

[0061] wherein N5 is an integer greater than or equal to 20; q = 1,..., n3+n4, is the serial number of running.

[0062] In a second aspect, the application provides a device for implementing the running training method based on a large model, which comprises:

[0063] a model pre-training module configured to pre-train a preset first large model;

[0064] a data recording module configured to record multi-modal data of an athlete;

[0065] an operation module configured to input the multi-modal data into the pre-trained first large model to obtain a running training scheme for the athlete;

[0066] wherein the pre-training of the preset first large model comprises:

[0067] recording historical multi-modal data of the athlete;

[0068] recording a training scheme corresponding to each piece of data in the historical multi-modal data;

[0069] importing the historical multi-modal data and the training scheme into the preset first large model to construct a fine-tuning data set of the running training scheme;

[0070] wherein the multi-modal data comprises one or a combination of the following: a running ability evaluation result, height, weight, a latest running performance score, gender, and age.

[0071] In a third aspect, the application provides a running training device based on a large model, which comprises a memory, a processor, and a user interface.

[0072] The memory is configured to store a computer program.

[0073] The user interface is configured to interact with a user.

[0074] The processor is configured to read the computer program in the memory, and when the processor executes the computer program, the running training method described above is implemented.

[0075] In a fourth aspect, the present application provides a processor-readable storage medium, characterized in that the processor-readable storage medium stores a computer program, and the processor executes the computer program to implement the above-mentioned large model-based running training method.

[0076] The present application has the following advantages:

[0077] The running training method and device provided by the present application record historical multi-modal data of athletes running and record training programs corresponding to each piece of multi-modal data, load a pre-trained large model, construct a fine-tuning data set for running training guidance, and the data set includes multi-modal data in various running situations and corresponding training guidance labels. Then, the multi-modal data of the athletes are recorded and input into the pre-trained large model to obtain corresponding running training programs, thereby providing personalized and real-time running training guidance and achieving scientific and efficient physical fitness improvement. BRIEF DESCRIPTION OF DRAWINGS

[0078] The accompanying drawings are included to provide a further understanding of embodiments of the application, and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and serve to explain the principles of the application. It is readily understood that the drawings are merely illustrative of typical embodiments of the application and therefore are not to be considered limiting of the scope of the application.

[0079] Figure 1 A flowchart of a large model-based running training method according to an embodiment of the present application;

[0080] Figure 2 A flowchart of a method for constructing a fine-tuning data set according to an embodiment of the present application;

[0081] Figure 3 A schematic diagram of a large model-based running training device according to an embodiment of the present application;

[0082] Figure 4 A schematic diagram of another large model-based running training device according to an embodiment of the present application.

[0083] Reference signs:

[0084] S1, step 1; S2, step 2; S201, step 201; S202, step 202; S203, step 203; 301, pre-training module; 302, data recording module; 303, operation module; 401, processor; 402, memory; 403, user interface. DETAILED DESCRIPTION

[0085] In the following, the technical solutions of the present application will be described clearly and completely in conjunction with the drawings, obviously, the described embodiments are some of the embodiments of the present application, rather than all the embodiments. It should be understood that these descriptions are only exemplary, and are not intended to limit the scope of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative labor should belong to the scope of protection of the present application.

[0086] In addition, in the following description, the description of well-known structures and techniques is omitted to avoid unnecessary confusion of the concepts disclosed in the present application.

[0087] In the description of the present application, it should be noted that, unless otherwise explicitly specified and limited, the orientation or positional relationship indicated by the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer" and the like is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present application and simplifying the description, and therefore cannot be understood as indicating or implying that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as limiting the present application. In addition, the terms "first", "second", "third" are only for description purposes, and cannot be understood as indicating or implying relative importance. The terms "mounting", "connecting", "connection" should be understood broadly, for example, it can be fixed connection, or detachable connection, or integrally connected; it can be mechanical connection, or electrical connection; it can be directly connected, or indirectly connected through intermediate medium, or the communication between two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0088] The exemplary embodiments will be described in detail herein, with examples shown in the drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of methods and systems consistent with some aspects of the present application, as detailed in the appended claims.

[0089] Current data applications mainly focus on data collection and basic physical fitness testing, and the guidance of after-school training is relatively lacking. Usually, it cannot be dynamically adjusted according to the individual physical fitness, training state and training effect of students. This leads to the homogeneity and inefficiency of training guidance, which cannot fully meet the needs of individualized training of students. In the training of professional athletes, individualized training methods also need to be developed according to the individual physical fitness, past training performance, etc. of the athletes, so as to provide more scientific training programs and improve sports performance. How to use large models to provide individualized and real-time training programs according to the physical test data, training images, training evaluation results and other multi-modal data of athletes is a technical problem that needs to be solved at present.

[0090] In view of the problems in the prior art, the present application provides a running training method and device based on a large model, which solves the problems of low automation, inaccurate target positioning and labeling, low work efficiency, high technical level requirement for test personnel, and tedious and time-consuming operation of the prior art.

[0091] Using the analysis and reasoning capabilities of large models and the processing capabilities of massive data can greatly improve the efficiency of training program production and improve the scientificity and rationality of training programs compared with manual analysis and processing. For example, manual analysis and processing of multi-modal data requires recording multi-modal data into a computer disk and producing a large number of intermediate analysis tables, classifying and organizing intermediate information and storing them as files, and matching with the height, weight, latest running score, gender, age, etc. of the athlete, and then finding the rules in the data through software program analysis, manual analysis and other methods. This analysis method requires storing a large number of intermediate files, consuming more computer storage resources, and if a special program is written for analysis, more storage space and computing resources will be consumed. Using the analysis and processing capabilities of large models on multi-modal data can not only save a lot of storage space for intermediate files, but also greatly improve the processing efficiency.

[0092] Method embodiment

[0093] In a first aspect, the present application provides a running training method based on a large model, as shown in Figure 1 Specifically, the method comprises steps S1 to S3:

[0094] S1, pre-training a preset first large model;

[0095] S2, record the multi-modal data of the athlete;

[0096] S3, input the multi-modal data into the pre-trained first large model to obtain the running training program of the athlete;

[0097] The process of pre-training the preset first large model is as shown in Figure 2As shown, comprising:

[0098] S201, record historical multi-modal data of the runner running;

[0099] S202, record the training scheme corresponding to each piece of data in the historical multi-modal data;

[0100] S203, import the historical multi-modal data and the training scheme into a preset first large model to construct a fine-tuning data set of the running training scheme;

[0101] The multi-modal data includes one or a combination of the following: running ability evaluation result, height, weight, latest running performance score, gender, and age.

[0102] As an optional example, the first large model is an InternVL2 large model. The first large model is fine-tuned by the historical multi-modal data, so that the running training scheme is more accurate. Specifically, the multi-modal data of the runner running can be recorded and collected by an intelligent running device, including the running ability evaluation in the historical training, height, weight, latest running performance score, gender, and age, and the training guidance corresponding to each piece of data is collected, a pre-trained InternVL2 model is loaded, a fine-tuning data set of the running training guidance is constructed, and the data set includes multi-modal data and corresponding training guidance scheme labels under various running situations. The InternVL2 model is fine-tuned by using the fine-tuning data set, so that the InternVL2 model learns how to generate personalized running training schemes according to the multi-modal data in the running process.

[0103] It should be noted that the historical multi-modal data in the embodiment refers to the multi-modal data recorded before S2 in the above embodiment, and the preset first large model is pre-trained using the historical multi-modal data. The multi-modal data in S2 refers to the multi-modal data of the current athlete, and the multi-modal data of the current athlete is input into the pre-trained and fine-tuned preset first large model to obtain the corresponding subsequent training scheme under the current movement state. For example, for a student, historical multi-modal data of N days before today has been collected, and the after-school training scheme corresponding to the multi-modal data of each day is collected, the historical multi-modal data and the corresponding after-school training scheme are input into the InternVL2 model for training to obtain a pre-trained InternVL2 model. For the student, the multi-modal data of today is collected and input into the pre-trained InternVL2 model to obtain the after-school running training scheme of today. For example, a student A's 100-meter sprint result today is 14 seconds, the running ability evaluation M1 of student A, height, weight, latest running result score, gender, age are input into the pre-trained InternVL2 model, and the after-school training scheme obtained today is high knee lift in place, and the step frequency of running is improved.

[0104] It should be noted that the training scheme corresponding to each piece of data in the historical multi-modal data can be a training scheme annotated by a professional training coach, that is, for a piece of multi-modal data, the training scheme corresponding to the multi-modal data is the best training scheme under the state of the athlete corresponding to the multi-modal data, and is a targeted training scheme designed for the problems existing in running or the direction that can be improved reflected by the multi-modal data.

[0105] It should be noted that the running training method of the application can be used for training of athletes of all ages, such as students, professional athletes, etc.

[0106] As an optional example, in the embodiment of the application, the running ability evaluation result is determined according to the following method:

[0107] The running ability evaluation result M1 is determined according to the following expression one:

[0108] M1 = "running ability evaluation:" + g1 + g2 + g3 + g4 + g5; (expression one)

[0109] Wherein, g1 is a posture standard evaluation result, used to indicate whether the posture of the runner conforms to the standard;

[0110] g2 is an explosive force evaluation result, used to indicate the running explosive force level of the athlete;

[0111] g3 is a stamina evaluation result, used to indicate the running stamina level of the athlete;

[0112] g4 is a running skill evaluation result, used to represent the running skill level of the athlete;

[0113] g5 is a training quality evaluation result, used to represent the running training level of the athlete.

[0114] Wherein, the running ability evaluation result M1 is string data, the + in the expression represents the splicing of the string, g1, g2, g3, g4 and g5 are string data. When splicing different strings, a separator can be added between the two spliced strings, for example, a comma “,”, or a semicolon “;”, or a space “”, or other special characters such as an asterisk “*”, a hash “#”, an exclamation mark “!” and the like, which are determined according to the needs. For example, if g1 = “perfect posture, efficient and labor-saving”, g2 = “powerful burst, has an advantage in sprinting”, g3 = “insufficient endurance, long-distance running ability needs to be improved”, g4 = “stable pace, good skill”, g5 = “training very frequently, training quality very high, training effect very high”, and the separator added in the middle is a semicolon “;”, then the result after processing by the above expression one is: M1 = “Running ability evaluation: perfect posture, efficient and labor-saving; powerful burst, has an advantage in sprinting; insufficient endurance, long-distance running ability needs to be improved; stable pace, good skill; training very frequently, training quality very high, training effect very high”.

[0115] In the embodiment of the present application, the running ability evaluation result is determined by the athlete's posture, power, endurance, skill and training quality.

[0116] As an optional example, the posture standard evaluation result g1 is determined according to the following method:

[0117] Pre-train a preset second large model using the athlete's historical running posture data;

[0118] Record the athlete's running posture images for consecutive N1 days, identify each running posture image through the pre-trained second large model, output whether the athlete's posture meets the standard, record the number of images n1 that meet the standard and the number of images n2 that do not meet the standard;

[0119] Determine the posture standard evaluation result g1 according to the following expression three:

[0120]

[0121] Wherein, ts1 is a set first judgment threshold, ts2 is a set second judgment threshold, ts3 is a set third judgment threshold, N1 is an integer greater than or equal to 20, for example, it can also be 30 days, 50 days, 60 days, etc.;

[0122] Pre-training the preset second large model using the historical running posture data of the athlete includes:

[0123] The historical running posture data of the athlete is recorded, and the historical running posture data is labeled to label the key posture elements of the athlete, including one or a combination of the following: trunk angle, arm swing angle, stride, and landing manner.

[0124] The historical running posture data is labeled and divided into standard and non-standard categories.

[0125] The preset second large model is pre-trained using the labeled historical running posture data, so that the preset second large model can recognize key posture features and associate the key posture features with "standard" or "non-standard" labels.

[0126] As an optional example, the preset second large model is the InternVL2 large model.

[0127] For example, the posture images of the student during the running process are recorded by the intelligent running device, and the image acquisition date is labeled to obtain the running posture images of the student for 30 consecutive days (the default highest option is evaluated when there is no record). The InternVL2 large model is fine-tuned to identify each running posture image and output whether the runner's posture is standard or not. The number of images n1 that meet the standard and the number of images n2 that do not meet the standard are recorded, and the posture standard evaluation is calculated according to the above expression three. For example, for an athlete, the posture standard evaluation result g1 is calculated as "posture standard, with room for improvement".

[0128] The fine-tuning process of the InternVL2 large model is as follows:

[0129] A large amount of running posture data (e.g., more than ten thousand) is collected, and the collected data is labeled to label the key posture elements of the runner, including trunk angle, arm swing angle, stride, landing manner, etc. At the same time, the data is divided into "standard" and "non-standard" categories. The collected standard and non-standard running posture data is used to fine-tune the multi-modal model, so that it can recognize and distinguish the subtle differences of running posture. During the fine-tuning process, the model is ensured to recognize key posture features and associate these features with "standard" or "non-standard" labels.

[0130] As an optional example, the explosive power evaluation result g2 is determined according to the following method:

[0131] The sprint project scores g2 of the athlete for N2 consecutive days are recorded i where i is the serial number of the sprint project score, i = 1, 2, 3,..., n3, and n3 is the number of sprint project scores of the athlete for N2 consecutive days.

[0132] The explosive power evaluation result g2 is determined according to the following Expression Four:

[0133]

[0134] wherein g 21 is the explosive power evaluation score, and g 21 is determined by the following Expression Five:

[0135]

[0136] wherein g21 i is the sub-zone score, and g21 i is determined by the following Expression Six:

[0137]

[0138] wherein N2 is an integer greater than or equal to 20.

[0139] For example, the performance scores of a student in each 50-meter, 50-meter*8, 60-meter, 100-meter, 200-meter, etc. sprint event are recorded by an intelligent running device, and the student's sprint performance for 30 consecutive days is obtained (the default highest option is evaluated when there is no record), and the explosive power evaluation result is determined according to the above-mentioned Expression Four. Assuming that the explosive power evaluation result of a certain athlete is g2 = "strong explosive power, very good at sprinting".

[0140] As an optional example, the endurance evaluation result g3 is determined according to the following method:

[0141] The long-distance running event scores g3 of an athlete for N3 consecutive days are recorded j wherein j is the serial number of the long-distance running event score, j = 1, 2, 3,..., n4, and n4 is the number of long-distance running event scores of the athlete for N3 consecutive days;

[0142] The endurance evaluation result g3 is determined according to the following Expression Seven:

[0143]

[0144] wherein g 31 is the endurance evaluation score and is determined by the following Expression Eight:

[0145]

[0146] wherein N3 is an integer greater than or equal to 20.

[0147] For example, the long-distance running results of a student for 30 consecutive days are recorded (the default highest option is evaluated when there is no record), and the endurance evaluation result g3 is determined using the above expression seven. It is assumed that the endurance evaluation result g3 of a student is "poor endurance, and fatigue after long-distance running".

[0148] As an optional example, the running skill evaluation result g4 is determined according to the following method:

[0149] The time t of each lap of a long-distance running event completed by an athlete in N4 days is recorded m , the total number of laps n5, the distance of each lap l1, m = 1,..., n5, and the number of completed laps.

[0150] When ,

[0151] g4 = pace stability, good skill;

[0152] When and and ,

[0153] g4 = pace control is very unstable, with large fluctuations, and lacks running skill;

[0154] In other cases:

[0155] g4 = pace fluctuation, and running skill needs to be further strengthened;

[0156] Where ts4 is a set fourth judgment threshold, ts5 is a set fifth judgment threshold, and ts6 is a set sixth judgment threshold.

[0157] Where N4 is an integer greater than or equal to 20; p = 1,..., n5, is the number of completed laps.

[0158] For example, the time t of each lap of a long-distance running event completed by an athlete for 400 meters, 800 meters, 1000 meters, 3000 meters, etc. is recorded m , and the total number of laps n5, the distance of each lap l1, m = 1,..., n5, and the number of completed laps (the default highest option is evaluated when there is no record). The running skill evaluation result is calculated according to the above steps. It is assumed that the running skill result of the athlete is g4 = "pace fluctuation, and running skill needs to be further strengthened".

[0159] As an optional example, the training quality evaluation result g5 is determined according to the following method:

[0160] The running results g5 of an athlete for N5 consecutive days are recorded p, p = 1,..., n3+n4, is the running serial number, the smaller the serial number, the earlier the time of the training record, n3 is the number of sprint event scores of the athlete in continuous N2 days, n4 is the number of long-distance running event scores of the athlete in continuous N3 days, wherein N2 is an integer greater than or equal to 20, and N3 is an integer greater than or equal to 20;

[0161] The training quality evaluation result g5 is determined according to the following expression nine:

[0162] g5 = g 51 + g 52 + g 53 ; (Expression nine)

[0163] In the expression nine, g 51 , g 52 and g 53 are concatenated, and the processing manner is the same as that of the expression one, which will not be described here.

[0164] In the expression nine, g 51 is the quantity evaluation, g 52 is the quality evaluation, g 53 is the effect evaluation, and satisfies the following expression ten, expression eleven, expression twelve and expression thirteen:

[0165]

[0166] In the expression nine, ts7 is the set seventh judgment threshold, ts8 is the set eighth judgment threshold, ts9 is the set ninth judgment threshold, ts 10 is the set tenth judgment threshold, ts 11 is the set eleventh judgment threshold, ts 12 is the set twelfth judgment threshold, ts 13 is the set thirteenth judgment threshold, ts 14 is the set fourteenth judgment threshold, ts 15 is the set fifteenth judgment threshold, ts 16 is the set sixteenth judgment threshold, and ts 17 is the set seventeenth judgment threshold.

[0167] In the expression nine, N5 is an integer greater than or equal to 20; q = 1,..., n3+n4, is the running serial number.

[0168] For example, the performance of each running training of a student is recorded by a smart running device, the running performances of the student for 60 consecutive days are taken, including the performances for 20 days of short-distance running and the performances for 40 days of long-distance running, and then the training quality evaluation result g5 of the student is calculated according to the above expression nine. It is assumed that the training quality evaluation result of the student is g5 = "training amount is relatively large # training quality is very high # training effect is relatively high".

[0169] In the embodiment of the application, the historical multi-modal data of the running of the athletes is recorded and collected, the training scheme corresponding to each piece of multi-modal data is recorded, a pre-trained large model is loaded, a fine-tuning data set for running training guidance is constructed, the data set includes multi-modal data in various running situations and corresponding training guidance labels. Then the multi-modal data of the athletes is recorded and input into the pre-trained large model to obtain the corresponding running training scheme, thereby providing personalized and real-time running training guidance and realizing scientific and efficient physical fitness improvement.

[0170] Device embodiment

[0171] In a second aspect, another specific embodiment of the application discloses a running training device based on a large model, as shown in the accompanying drawings, which comprises: Figure 3

[0172] The model pre-training module 301 is configured to pre-train a preset first large model;

[0173] The data recording module 302 is configured to record the multi-modal data of the athletes;

[0174] The operation module 303 is configured to input the multi-modal data into the pre-trained first large model to obtain the running training scheme of the athletes;

[0175] The pre-training of the preset first large model comprises:

[0176] The historical multi-modal data of the running of the athletes is recorded;

[0177] The training scheme corresponding to each piece of data in the historical multi-modal data is recorded;

[0178] The historical multi-modal data and the training scheme are imported into the preset first large model to construct a fine-tuning data set for the running training scheme;

[0179] The multi-modal data comprises one or a combination of the following: a running ability evaluation result, a height, a weight, a latest running performance score, a gender, and an age.

[0180] As an optional example, the operation module 303 is further configured to determine the running ability evaluation result M1 according to the following method:

[0181] ​The running ability evaluation result M1 is determined according to the following expression:

[0182] M1 = "Running ability evaluation:" + g1 + g2 + g3 + g4 + g5;

[0183] wherein g1 is a posture standard evaluation result, used to indicate whether the posture of the runner conforms to the standard;

[0184] g2 is an explosive force evaluation result, used to indicate the running explosive force level of the athlete;

[0185] g3 is a stamina evaluation result, used to indicate the running stamina level of the athlete;

[0186] g4 is a running skill evaluation result, used to indicate the running skill level of the athlete;

[0187] g5 is a training quality evaluation result, used to indicate the running training level of the athlete;

[0188] wherein the running ability evaluation result M1 is string data, the + in the expression indicates string splicing, and g1, g2, g3, g4 and g5 are string data.

[0189] As an optional example, the operation module 303 is further configured to determine the posture standard evaluation result g1 according to the following method:

[0190] Pre-train a preset second large model using historical running posture data of the athlete;

[0191] Record the running posture images of the athlete for N1 consecutive days, identify each running posture image through the pre-trained second large model, output whether the posture of the runner conforms to the standard, record the number n1 of images conforming to the standard and the number n2 of images not conforming to the standard;

[0192] Determine the posture standard evaluation result g1 according to the following expression:

[0193]

[0194] wherein ts1 is a set first judgment threshold, ts2 is a set second judgment threshold, ts3 is a set third judgment threshold, and N1 is an integer greater than or equal to 20;

[0195] Pre-training the preset second large model using historical running posture data of the athlete includes:

[0196] Record the historical running posture data of the athlete, label the historical running posture data, and label the key posture elements of the athlete, including one or a combination of the following: trunk angle, arm swing angle, stride, and landing method;

[0197] The historical running posture data is labeled and divided into a standard class and a non-standard class;

[0198] The preset second large model is pre-trained using the labeled historical running posture data, so that the preset second large model can identify key posture features and associate the key posture features with "standard" or "non-standard" labels.

[0199] As an optional example, the operation module 303 is further configured to determine the explosive power evaluation result g2 according to the following method:

[0200] Record the sprint item scores g2 of the athlete for consecutive N2 days i , wherein i is the serial number of the sprint item score, i = 1, 2, 3, …, n3, and n3 is the number of sprint item scores of the athlete for consecutive N2 days;

[0201] The explosive power evaluation result g2 is determined according to the following expression:

[0202]

[0203] , wherein g 21 is the explosive power evaluation score, and g 21 is determined by the following expression:

[0204]

[0205] , wherein g21 i is the partition score, and g21 i is determined by the following expression:

[0206]

[0207] , wherein N2 is an integer greater than or equal to 20.

[0208] As an optional example, the operation module 303 is further configured to determine the endurance evaluation result g3 according to the following method:

[0209] Record the long-distance running item scores g3 of the athlete for consecutive N3 days j , wherein j is the serial number of the long-distance running item score, j = 1, 2, 3, …, n4, and n4 is the number of long-distance running item scores of the athlete for consecutive N3 days;

[0210] The endurance evaluation result g3 is determined according to the following expression:

[0211]

[0212] , wherein g 31 is the endurance evaluation score and is determined by the following expression:

[0213]

[0214] wherein N3 is an integer greater than or equal to 20;

[0215] As an optional example, the operation module 303 is further configured to determine the running skill evaluation result g4 according to the following method:

[0216] record the time t of the athlete to complete each lap in the long-distance running event in N4 days m , the corresponding total number of laps n5, the distance of each lap l1, m = 1,..., n5,..., which is the serial number of the completed laps;

[0217]

[0218] g4 = pace stability, good skill;

[0219] when and and :

[0220] g4 = pace control is very unstable, large fluctuations, lack of running skills;

[0221] in other cases:

[0222] g4 = pace fluctuation, running skills need to be further strengthened;

[0223] wherein ts4 is a set fourth judgment threshold, ts5 is a set fifth judgment threshold, and ts6 is a set sixth judgment threshold;

[0224] wherein N4 is an integer greater than or equal to 20; p = 1,..., n5, which is the serial number of the completed laps.

[0225] As an optional example, the operation module 303 is further configured to determine the training quality evaluation result g5 according to the following method:

[0226] record the running results g5 of the athlete for consecutive N5 days p , p = 1,..., n3 + n4, which is the serial number of the running, the smaller the serial number, the earlier the training record time, n3 is the number of scores of the athlete in the short-distance running event for consecutive N2 days, and n4 is the number of scores of the athlete in the long-distance running event for consecutive N3 days, wherein N2 is an integer greater than or equal to 20, and N3 is an integer greater than or equal to 20;

[0227] The training quality evaluation result g5 is determined according to the following expression:

[0228] g5 = g 51 + g 52 + g53 wherein the + sign represents string concatenation;

[0229] wherein, g 51 is the quantity evaluation, g 52 is the quality evaluation, g 53 is the effect evaluation, and satisfies the following expression:

[0230]

[0231]

[0232] wherein, ts7 is a set seventh judgment threshold, ts8 is a set eighth judgment threshold, ts9 is a set ninth judgment threshold, ts 10 is a set tenth judgment threshold, ts 11 is a set eleventh judgment threshold, ts 12 is a set twelfth judgment threshold, ts 13 is a set thirteenth judgment threshold, ts 14 is a set fourteenth judgment threshold, ts 15 is a set fifteenth judgment threshold, ts 16 is a set sixteenth judgment threshold, ts 17 is a set seventeenth judgment threshold.

[0233] wherein, N5 is an integer greater than or equal to 20; q = 1,..., n3 + n4, is the serial number of running.

[0234] It should be noted that the running training device provided in the second aspect and the running training method provided in the first aspect belong to the same inventive concept, solve the same technical problem, and obtain the same technical effect, and thus will not be described here.

[0235] In a third aspect, the present application proposes a running training device based on a large model, as shown in Figure 4 , which comprises a memory and one or more processors.

[0236] The memory stores one or more application programs, and the one or more application programs are adapted to be executed by the one or more processors to implement the running training device method of the first aspect.

[0237] As shown in Figure 4 , the electronic device comprises a processor 401 and a memory 402. The processor 401 and the memory 402 are connected, such as through a bus interface.

[0238] The structure of the electronic device does not constitute a limitation on the embodiments of the present application.

[0239] The processor 401 can be a CPU, a general-purpose processor, a DSP, an ASIC, an FPGA, or other programmable logic device, transistor logic device, hardware component, or any combination thereof. It can implement or execute various exemplary logical blocks, modules, and circuits described in conjunction with the present disclosure. The processor 401 can also be a combination of computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, and the like.

[0240] The bus interface can include a path for transmitting information between the above-mentioned components. The bus interface can be a PCI bus or an EISA bus, etc. The bus interface can be divided into an address bus, a data bus, a control bus, etc. For convenience of representation, Figure 4 In the figure, only one thick line is used to represent, but it does not mean that there is only one bus or one type of bus.

[0241] The memory 402 can be a ROM or other type of static storage device that can store static information and instructions, a RAM or other type of dynamic storage device that can store information and instructions, an EEPROM, a CD-ROM or other optical disk storage, an optical disk storage (including a compact disk, a laser disk, an optical disk, a digital versatile disk, a Blu-ray disk, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and capable of being accessed by a computer, but not limited thereto.

[0242] In a fourth aspect, the present application provides a computer-readable storage medium having stored thereon a computer program capable of being loaded and executed by a processor to implement the running training method of the first aspect.

[0243] The applicant of the present application has made a detailed description and explanation of the embodiments of the present application in combination with the drawings of the specification, but those skilled in the art should understand that the above embodiments are only preferred embodiments of the present application, and the detailed description is only to help the reader better understand the spirit of the present application, and is not a limitation on the protection scope of the present application, on the contrary, any improvement or modification based on the spirit of the present application should fall within the protection scope of the present application.

[0244] It should be pointed out finally that the above embodiments are only used to illustrate the technical solutions of the embodiments of the present application, but not to limit them. Although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those skilled in the art that the technical solutions recorded in the foregoing embodiments can be modified, or some technical features can be replaced equivalently; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application. Any changes or replacements that can be easily thought of by those skilled in the art within the technical scope disclosed by the present application should be covered within the protection scope of the present application.

Claims

1. A method for running training based on a large model, characterized by, The method comprises the following steps: S1, pre-training a preset first large model; S2, recording multi-modal data of an athlete; S3, inputting the multi-modal data into the pre-trained first large model to obtain a running training plan for the athlete; wherein the pre-training of the preset first large model comprises: recording historical multi-modal data of the athlete running; recording a training plan corresponding to each piece of data in the historical multi-modal data; importing the historical multi-modal data and the training plan into the preset first large model to construct a fine-tuning data set of the running training plan; wherein the multi-modal data comprises one or a combination of the following: running ability evaluation result, height, weight, latest running performance score, gender, and age; wherein the preset first large model is an InternVL2 large model.

2. A method of running training according to claim 1, wherein, The running ability evaluation result is determined according to the following method: The running ability evaluation result M1 is determined according to the following expression: M1 = "running ability evaluation:" + g1 + g2 + g3 + g4 + g5; wherein g1 is a posture standard evaluation result; g2 is an explosive strength evaluation result; g3 is an endurance evaluation result; g4 is a running skill evaluation result; g5 is a training quality evaluation result; wherein the running ability evaluation result M1 is string data, the "+" in the expression represents string splicing, and g1, g2, g3, g4, and g5 are string data.

3. A method of running training according to claim 2, wherein, The posture standard evaluation result g1 is determined according to the following method: pre-training a preset second large model using historical running posture data of the athlete; recording running posture images of the athlete for N1 consecutive days, identifying each running posture image through the pre-trained second large model, outputting whether the posture of the runner conforms to the standard, and recording the number n1 of images conforming to the standard and the number n2 of images not conforming to the standard; determining the posture standard evaluation result g1 according to the following expression: wherein ts1 is a preset first judgment threshold, ts2 is a preset second judgment threshold, ts3 is a preset third judgment threshold, and N1 is an integer greater than or equal to 20; The pre-training of the preset second large model using the historical running posture data of the athlete comprises: recording the historical running posture data of the athlete, labeling the historical running posture data, and labeling the key posture elements of the athlete, which include one or a combination of the following: trunk angle, arm swing angle, stride, and landing method; labeling the historical running posture data and dividing it into standard and non-standard classes; pre-training the preset second large model using the labeled historical running posture data, so that the preset second large model can identify key posture features and associate the key posture features with "standard" or "non-standard" labels; wherein the preset second large model is an InternVL2 large model.

4. The method of claim 2, wherein, The explosive strength evaluation result g2 is determined according to the following method: record the sprint item scores g2i of the athlete for N2 consecutive days, where i is the serial number of the sprint item score, i = 1, 2, 3, …, n3, and n3 is the number of sprint item scores of the athlete for N2 consecutive days; determine the explosive force evaluation result g2 according to the following expression: wherein g 21 is the burst force evaluation score, and g 21 is determined from the following expression: where g21 i is the zone score, and g21 i is determined from the following expression: where N2 is an integer greater than or equal to 20.

5. The method of claim 2, wherein, The endurance evaluation result g3 is determined according to the following method: record the long-distance running item scores g3 of the athlete for consecutive N3 days j wherein j is the serial number of the long-distance running item score, j = 1, 2, 3, …, n4, n4 is the number of long-distance running item scores of the athlete for consecutive N3 days; determine the endurance evaluation result g3 according to the following expression: where g 31 is the resistance evaluation score and is determined by the following equation: where N3 is an integer greater than or equal to 20.

6. The method of claim 2, wherein, The running skill evaluation result g4 is determined according to the following method: record the time t of each lap completed by the athlete in the long run event in n4 days m corresponding total number of laps n5, distance per lap l1, m = 1,..., n5, number of laps completed When Time: g4 = pace stability, good skill; When and and when: g4 = pace control is very unstable and fluctuates greatly, lacking running skill; otherwise: g4 = pace fluctuates, running skill needs to be further strengthened; where ts4 is a set fourth judgment threshold, ts5 is a set fifth judgment threshold, and ts6 is a set sixth judgment threshold; where N4 is an integer greater than or equal to 20; p = 1, …, n5, is the serial number of the completed laps.

7. A large model-based running training device, characterized by, The device for implementing the running training method of any one of claims 1 to 6 comprises: a model pre-training module configured to pre-train a preset first large model; a data recording module configured to record multi-modal data of an athlete; an operation module configured to input the multi-modal data into the pre-trained first large model to obtain a running training plan for the athlete; wherein the pre-training of the preset first large model comprises: recording historical multi-modal data of an athlete running; recording a training plan corresponding to each piece of data in the historical multi-modal data; importing the historical multi-modal data and the training plan into a preset first large model to construct a fine-tuning data set for a running training plan; wherein the multi-modal data includes one or a combination of the following: running ability evaluation result, height, weight, latest running performance score, gender, and age; wherein the preset first large model is an InternVL2 large model.

8. A large model-based running training device, characterized by, comprises a memory, a processor, and a user interface; the memory is used to store a computer program; the user interface is used to interact with a user; the processor is used to read the computer program in the memory, and the processor implements the running training method of any one of claims 1 to 6 when executing the computer program.

9. A processor-readable storage medium, comprising: The processor readable storage medium stores a computer program, and the processor implements the running training method based on the large model of any one of claims 1 to 6 when executing the computer program. The processor readable storage medium stores a computer program, and the processor implements the running training method based on the large model of any one of claims 1 to 6 when executing the computer program.

Images