8 results about "Cross-training" patented technology

The invention provides a transfer learning fatigue life calculation method and system, and the method comprises the steps: obtaining training data, and determining key feature information corresponding to a key cycle number; constructing a training data set, taking the key feature information as a source domain data set, and dividing the target domain data set into the training data set and a test data set; establishing a main model and an auxiliary model, and sequentially carrying out coupling cross training on the main model and the auxiliary model by utilizing the source domain data set and the target domain data set; and calculating a difference value between the prediction results of the main model and the auxiliary model, and outputting a minimum training model and a corresponding prediction result when the difference value is smaller than a target difference value. According to the method, the comprehensive model of the main model and the auxiliary model is established, the comprehensive model is high in prediction precision, effectively prevents excessive generalization and is suitable for fatigue life calculation under the small sample condition, the main model and the auxiliary model are used for cross training, the training prediction difference value of the main model and the auxiliary model is minimum as the target, repeated iteration training is carried out, and fatigue life calculation is more accurate. And an optimal solution can be obtained.

The invention discloses a full-view prompt generation and wrong case discrimination cross training model for aspect-level emotion triple extraction and a model training method. The invention discloses a full view prompt generation and wrong case discrimination cross training model for aspect-level emotion triple extraction. The full view prompt generation and wrong case discrimination cross training model comprises a generation model input embedding module, an emotion triple generation model, a discrimination model input embedding module, an emotion triple discrimination model and an emotion triple output module. According to the method, the generation model is guided to learn semantic information in the label emotion triples from different perspectives through label sequence prompt of the full view, generation error example characteristics of the generation model are learned through the discrimination model, and extraction results of the full view are discriminated and filtered so as to eliminate the emotion triples with the adjacency dependence mismatch problem, so that the efficiency of the whole view is improved. Meanwhile, generation errors caused by multi-view prediction convergence are avoided; through the alternative training collaborative optimization scheme of the generative model and the discriminant model, the precision of the model for integrally extracting the emotion triad is improved.

A method of compiling workout information in a workout session includes, at a workout server, receiving, via a network, first login information from a first exercise device; receiving, via the network, first workout information from the first exercise device; recording the first workout information in an active workout session; receiving, via the network, second login information from a second exercise device; receiving, via the network, second workout information from the second exercise device; and recording the second workout information in the active workout session.

The utility model discloses an adjustable cross fork training device which comprises a height adjusting assembly, the lower end of a fixing frame is fixedly arranged on a base, a sliding seat is arranged at the upper end of the fixing frame in a sliding mode, the height adjusting assembly is arranged between the fixing frame and the sliding seat and used for driving the sliding seat to slide, and the sliding seat is connected with the upper end of a circulating pull rod. The lower end of the circulating pull rod is connected with the seat, the seat is slidably arranged on a guide rod, the guide rod is fixedly arranged on the base, adjusting leg rods are hinged to the two sides of the seat, the lower ends of the adjusting leg rods are slidably arranged on the base, and the height adjusting assembly is electrically connected with the control center. And the leg rods can be adjusted to extend and retract under the action of the control center. The height adjusting assembly drives the seat to move up and down, so that cross-fork trainers with different heights and different states can be exercised, and the cross-fork trainers can automatically adjust the cross-fork trainers according to needs without assistance of other people.

A training apparatus having a progress indicator and unified base for different apparatus types, such as for an ergometer, speed bike, recumbent bike, home training bike, rowing apparatus, elliptical trainer, cross-trainer or treadmill. The fitness apparatus of every apparatus type is based on a structure which is unified in terms of technical design, according to which the frame or pedestal of the fitness apparatus has rollers which are arranged on the frame or pedestal such that a tilt-stable positioning of the frame with respect to the subsurface and a longitudinal displacement of the frame or of the frame together with the pedestal relative to the subsurface is enabled by means of an actuator depending on the measure of success of the user training.

The present application relates to the technical field of deep learning, in particular to a novel deep learning method and device based on symmetric cross compound training, which specifically comprises: obtaining first sample data, dividing the first sample data into a first training set, a first validation set and a first test set, and further dividing the first training set into a first data set and a second data set; taking the first data set and the second data set alternately as a training set and a validation set to perform symmetric cross training, obtaining outlier samples and true samples; performing compound training on the outlier samples and the true samples, simultaneously performing output correction classification, and obtaining a first optimal model. The present application changes the traditional deep learning training method mode, uses the function of deep learning feature extraction to automatically identify and separate the outlier samples in the training set, and further improves the model performance.

The utility model relates to the technical field of rehabilitation training equipment, in particular to a handle steering quick-changing mechanism of upper and lower limb active and passive rehabilitation training equipment, which is applied to a lower limb trainer, an upper limb trainer is arranged at the top of the lower limb trainer, the upper limb trainer comprises a shell, a mounting seat is arranged in the shell, and a handle is arranged on the mounting seat. A motor is arranged at the top of the mounting base, a first flange plate is arranged on one side of the motor, a first protective sleeve is arranged on one side of the first flange plate, a left handle is arranged on one side of the first protective sleeve, and a second flange plate is arranged on the side, away from the first flange plate, of the motor. According to the handle steering quick-changing mechanism of the upper and lower limb active and passive rehabilitation training equipment, the adjusting mechanism is arranged, the crank can be quickly reversed by pulling up a reset spring positioning piece in the axial direction and then rotating the crank, and therefore the crank reversing efficiency and operation convenience are greatly improved; switching of two training modes of vertical cross training and vertical horizontal training is achieved.

The present specification relates to the technical field of petrochemical production, and particularly relates to a three-agent use data comprehensive analysis and optimization method, device and equipment. Including: obtaining historical production data of three-agent related materials in a production line, classifying the obtained historical production data according to the categories of the three agents; then establishing the historical production data into multiple data models according to the categories, wherein the data models include: a price ratio model, a cost model and a device model; the data models are subjected to multi-condition constraints according to actual conditions, and an overall optimization model is generated through cross-training, the production demand in actual production is input into the optimization model to generate optimization suggestions, so that the operating personnel can purchase, produce and other processes according to the optimization suggestions. Thus, an optimization model capable of generating an optimization suggestion covering the whole process of three-agent production is realized.