34results about How to "Improve tuning efficiency" patented technology

The invention relates to the technical field of large models, and discloses a large model adjustment method, device and equipment and a storage medium, the method comprises the following steps: carrying out segmentation deployment on a to-be-adjusted large model to obtain sub-models deployed at different end side equipment, each sub-model comprising different numbers of coding and decoding layers and adjustable adapters, a corresponding adjustable adapter is embedded behind each coding and decoding layer; updating the adjustable adapter layer by layer according to the training state of the to-be-adjusted large model to obtain a large model of the adjustable adapter embedded with updated configuration; and performing forward propagation calculation and backward propagation calculation on the large model of the adjustable adapter embedded with the updated configuration to obtain a large model after parameter adjustment. The parameter adjustment efficiency of the large model can be effectively improved, and equipment resources are saved.

The invention discloses a listed company risk assessment method based on cross validation and Bayesian optimization, and the method comprises the following steps: 1, obtaining the comprehensive factor data of a listed company, carrying out the dimension reduction of the factor data through employing a principal component analysis method, obtaining a principal component factor, carrying out the effectiveness verification of the principal component factor through an IC and IR analysis method, and carrying out the calculation of the risk of the listed company; selecting a plurality of principal component factors as input features; 2, on the basis of the multi-factor model, constructing an ensemble learning index enhancement model by taking a decision tree and an AdaBoost algorithm as cores; 3, performing system optimization on parameters of the ensemble learning index enhancement model by adopting cross validation and a Bayesian method, and finally determining an optimal parameter combination to obtain a trained ensemble learning index enhancement model; and 4, evaluating the risk of the listed company by using the trained integrated learning index enhancement model. According to the method, the calculation resource consumption of the model can be reduced, the calculation efficiency is improved, and the over-fitting risk of the model can be reduced.

The invention provides a parameter optimization method and device, a text generation method and device, equipment and a medium, and relates to the field of artificial intelligence, and the method comprises the steps: obtaining a question text, a baseline answer text corresponding to the question text, a pre-training language model and a parameter pool; wherein the parameter pool comprises a plurality of groups of sampling parameters; aiming at each group of sampling parameters in the parameter pool, inputting the question text into a pre-training language model which is provided with the sampling parameters to obtain a generative answer text, and determining the fitness of the sampling parameters according to the similarity between the generative answer text and the baseline answer text; carrying out sampling parameter updating on the parameter pool by utilizing a genetic algorithm and the fitness; if the parameter pool does not meet the preset update stop condition, entering a step of inputting the question text into a pre-training language model with set sampling parameters to obtain a generative answer text; if the parameter pool meets the preset update stop condition, setting the sampling parameter with the optimal fitness as the optimal sampling parameter; and the sampling parameter tuning efficiency can be improved.

The invention discloses a database parameter adjusting and optimizing method based on parameter type customization optimization, and belongs to the field of database configuration optimizing.The method comprises the steps that a database parameter space is divided into a continuous type parameter subspace and a classified type parameter subspace according to parameter types, and the continuous type parameter subspace and the classified type parameter subspace are projected to a low-dimensional space; a GP model and an SMAC model are used as proxy models in the low-dimensional continuous parametric subspace and the classification parametric subspace respectively to search continuous parametric configuration and classification parametric configuration; interaction between the GP model and the SMAC model is realized through iterative optimization and a context-based communication mechanism; and mapping a parameter configuration result back to a database parameter space for performance evaluation. According to the method, the database parameter tuning performance and efficiency can be effectively improved, and a user is allowed to search database parameter configuration with higher performance in a shorter time.

The invention discloses an adjusting and optimizing method and system based on a large power dispatching model, and belongs to the technical field of power dispatching, and the method comprises the following steps: obtaining historical power dispatching knowledge, and dividing the historical power dispatching knowledge into regional knowledge and global knowledge based on regional characteristics; constructing a first teacher model based on the difference between the global knowledge and the regional knowledge, and constructing a student model based on the lowest fluctuation cost; obtaining regional scheduling correlation according to historical power scheduling knowledge, and constructing a second teacher model according to the regional scheduling correlation; associating the first teacher model, the second teacher model and the student model based on knowledge distillation, and constructing a large power dispatching model; and outputting a tuning strategy according to the large power dispatching model and the current power data. The beneficial effects of the invention are that the output of the power tuning strategy gives consideration to pertinence and efficiency.

The invention relates to a design method of a surgical robot controller, the controller and a surgical robot. The design method comprises the steps that a mathematical model is established according to state parameters of the tail end of the mechanical arm; and inputting preset specific instruction data into the surgical robot, and obtaining motion data of the motion of the mechanical arm, so as to fit the mathematical model. And inputting a target test position, and calculating a test difference value between the position of the tail end of the mechanical arm and the target test position. And the test difference value is decomposed into moving distances of the tail end of the mechanical arm in different directions. And setting a control logic to adjust the movement of the mechanical arm so as to keep the movement of the tail end of the mechanical arm in different directions and simultaneously reach the end points in all directions. And control parameters for controlling the mechanical arm are determined through a parameter setting tool. And testing the surgical robot, sampling motion data, and comparing the motion data with simulation data obtained by the mathematical model to determine whether to adjust the selected control parameters or not.

The invention discloses a full-automatic quantitative optimization method for a mixed precision model, manual intervention is reduced and the optimization efficiency is improved through the processes of optimization of the mixed precision model through a full-precision model and automatic quantitative optimization of the mixed precision, the model is optimized through an iteration strategy and random deletion and recovery layers, the optimal configuration can be gradually approached, and the model precision is improved. And based on an automatic evaluation and decision-making mechanism of a PR key index, the method flexibly adapts to different models and calibration tables, has wide applicability, provides traceability through detailed log records, and is convenient to analyze and reproduce the tuning process.

The invention discloses a parameter tuning method and device of an operating system and electronic equipment. The method relates to the technical field of artificial intelligence, and comprises the following steps: obtaining demand information input by a user; the demand information is input into a target model, a parameter configuration file and a test command are output, the target model is a model obtained by training a large language model through a parameter corpus of the operating system, and the parameter configuration file comprises parameter information for parameter configuration of the target operating system; the test command is used for testing the running performance of the target operating system; and according to the parameter configuration file and the test command, performing parameter tuning processing on the target operating system. Through the method and the device, the problem of relatively low tuning efficiency of the operating system caused by relatively long time consumption for parameter tuning of the operating system based on the Bayesian theory in related technologies is solved.

The invention provides a multi-lithology identification and prediction method based on a semi-supervised learning framework, and the method comprises the steps: generating a data set, and carrying out the preprocessing of the data set; obtaining a basic model, and initializing parameters of the basic model; optimizing parameters of the basic model based on a Bayesian function; performing iterative training on the basic model through a variational EM algorithm; the variational EM algorithm comprises a variational expectation step and a variational maximization step; whether the variational EM algorithm converges is judged, then the classification accuracy of the model is evaluated, and a prediction result of data without labels is observed. According to the method, an innovative solution is provided for multi-lithology identification and prediction, and the method has important practical significance and application value in the field of mineral exploration.

The utility model discloses a four -axis rotary table regulation structure relates to four -axis rotary table regulation technical field, and its technical scheme is: including lathe table, the bottom plate is equipped on lathe table top, and lathe table is connected with the bottom plate through bolt fixing, and the sliding slot is seted up in bottom plate one side, and the sliding block is equipped in the inside of sliding slot, and the fixed installation plate is equipped in the top of sliding block, and the rotary table body is equipped on the top of installation plate, and two adjusting assemblies are equipped on the top of installation plate, and the positioning assembly is equipped on the top of bottom plate, adjusting assembly includes adjusting block, and the beneficial effect is: the utility model discloses the cooperation of six -sided screw rod and adjusting block through the setting adjusting assembly, can realize the accurate adjustment of rotary table body position, and the position of adjusting block can be fine -tuned to the rotation six -sided screw rod, thereby accurately adjusting the position of rotary table body, ensures that rotary table can reach the ideal position accuracy, greatly improves the machining accuracy of lathe, satisfies the demand of high -precision machining.

The present invention relates to the field of optical lens technology, and specifically to an optical lens calibration method and system. First, an original optical lens parameter vector is obtained, and an image quality evaluation value of the original parameter vector is calculated by a ray tracing method, so as to effectively distinguish the optical lens that needs to be calibrated and determine the first optical lens parameter vector; then, the first optical lens parameter vector is input into a conditional generative adversarial network to generate a second optical lens parameter vector, and the conditional generative adversarial network introduces image quality evaluation loss and manufacturing constraint loss, which can guide the generator to generate an optical lens parameter vector that meets imaging requirements and complies with physical manufacturing limitations; then, the image quality evaluation value of the second optical lens parameter vector is calculated by a ray tracing method to ensure that the second optical lens parameter vector can meet the imaging requirements, and then the optical lens is calibrated according to the second optical lens parameter vector, thereby improving the accuracy of optical lens calibration and reducing the calibration cost.

The invention relates to the technical field of GPU power consumption control, in particular to an automatic tuning system for GPU power consumption control parameters, which comprises a test module, a GPU chip, a processor and a memory in which a computer program is stored, and when the computer program is executed by the processor, the following steps are realized: acquiring parameter definition information and configuring the parameter definition information into the GPU chip by using the test module, forming a plurality of test samples according to the plurality of temporary values respectively corresponding to all the variable parameters, respectively configuring each test sample into the GPU chip by using a test module, obtaining a corresponding performance power consumption ratio, and determining the test sample corresponding to the maximum performance power consumption ratio as a target sample, the GPU power consumption control parameters are automatically adjusted and optimized through the test module, manpower resource consumption is greatly reduced, the utilization rate of idle time can be increased, and therefore the GPU power consumption control parameter adjusting and optimizing efficiency is improved.

The invention relates to the technical field of micro-nano optics, in particular to a polarization beam splitter parameter tuning method and system based on a micro-nano grating, and the method comprises the steps: obtaining test optical characteristic parameters of a polarization beam splitter to be evaluated, and judging whether the parameter tuning needs to be carried out on the polarization beam splitter to be evaluated or not according to the test optical characteristic parameters and preset optical characteristic parameters; if parameter tuning needs to be carried out on the to-be-evaluated polarization beam splitter, to-be-tuned optical characteristic parameters of the to-be-tuned polarization beam splitter are acquired; determining singular structure characteristic parameters of the polarization beam splitter needing parameter tuning according to the optical characteristic parameters to be tuned; and carrying out optimization processing on the singular structure characteristic parameters of the polarization beam splitter needing parameter tuning. According to the invention, the parameter tuning efficiency of the polarization beam splitter is effectively improved, the structural factors influencing the performance are accurately positioned, and finally the optical performance of the polarization beam splitter is ensured to meet the preset requirements.

The application discloses a CDC shock absorber whole vehicle abnormal sound quantification method, device, equipment and medium, relates to the automobile NVH test and diagnosis technical field, and the method comprises the following steps: obtaining subjective evaluation based on abnormal sound perception, vibration signals and sound signals measured by sensors under different driving conditions; obtaining vibration characteristic values based on the vibration signals, obtaining sound characteristic values based on the sound signals, and converting the subjective evaluation into standardized subjective score values; calculating standardized vibration values and standardized sound values, and calculating an abnormal sound comprehensive quantification index in combination with the standardized subjective score values; dividing the abnormal sound levels based on the abnormal sound comprehensive quantification index, combining the divided abnormal sound levels and the vibration characteristic values, and outputting adjustment suggestions for the CDC shock absorber. The application realizes objective and accurate quantification of the abnormal sound degree through multi-source data fusion and characteristic frequency analysis.

The utility model discloses an upper supporting mechanism of a packaging box cover plate forming machine. The upper supporting mechanism comprises a feeding conveying belt and a material taking device. The feeding conveying belt conveys inner supports one by one. The material taking device is in transmission connection with a transmission device for driving the material taking device to take and discharge materials, and the material taking device comprises a material sucking unit and an adjusting driving assembly; the material suction unit comprises a first material suction piece and a second material suction piece, and the first material suction piece and the second material suction piece take materials respectively. The adjusting driving assembly is in transmission connection with the first suction part and the second suction part and drives the first suction part and the second suction part to get away from each other or get close to each other. In the material taking state, the positions of the first material suction piece and the second material suction piece are matched with the position of the inner support on the feeding conveying belt; and in the upward supporting state, the first suction piece and the second suction piece are matched with the specification of the bottom plate in position. The adjustment driving assembly rapidly adjusts the distance between the first suction piece and the second suction piece according to requirements, the adaptability of the equipment is effectively improved, stable operation of the equipment is guaranteed, and adjustment is convenient.

The invention relates to the technical field of camera debugging, in particular to a camera module tuning method, and aims to solve the technical problem of how to realize tuning of a camera module more efficiently. In order to achieve the purpose, the collaborative interaction system is created, and the collaborative interaction system comprises a local code bin, an automatic file verification module and a temporary storage area. In response to the adjustment and optimization service provider submitting adjustment and optimization data to the local code warehouse, triggering an automatic file verification module to execute automatic verification; if the automatic verification is passed, submitting the tuning data to a temporary storage area and generating a notification, otherwise, returning the tuning data to a local code bin and prompting a failure reason; in response to the notification, pulling the tuning data to a local bin for the personnel of the tuning demand side to perform local verification; if the local verification is passed, the tuning data is combined into the code general warehouse of the tuning demand side, otherwise, the problem feedback information is transmitted back to the local code warehouse for further tuning of the tuning demand side, the tuning efficiency is improved, and the communication cost is reduced.

The invention relates to the technical field of space remote sensing camera in-orbit adjustment, in particular to a space remote sensing camera in-orbit adjustment and optimization method, which comprises the following steps of: firstly, testing in a static state after ensuring that the attitude of a space remote sensing camera is stable based on an area array working mode of a time delay integral linear array detector, and roughly adjusting the position of a secondary mirror through star observation; when the size and the shape of the obtained defocused spot are close to the size of a circular defocused spot and reach the minimum size, rough adjustment of a focusing position is started, and then fine adjustment of a secondary mirror position and the focusing position is carried out based on a blade edge target on the ground; and finally, the time delay integral linear array detector is tested in a push-broom state, and fine in-orbit tuning is carried out on the premise that it is confirmed that the TDI direction and the image motion speed are correct. According to the method, by setting the reasonable tuning steps, the tuning efficiency can be effectively improved, so that the tuning time is saved, and the consumption of on-orbit resources is reduced.

The utility model relates to laser cutting technical field especially is a kind of glass laser cutting device, including laser cutting mechanism and moving platform, the laser cutting mechanism includes laser, reflecting mirror piece and objective, the moving platform is located the below of objective, the moving platform is used to place glass sheet, glass sheet includes the upper layer, intermediate layer and lower layer of primary connection, the laser is used to generate laser, and laser is cut glass sheet after passing reflecting mirror piece and objective.The utility model provides a kind of glass laser cutting device, and different cutting depth can effectively improve product yield to promote;Different cutting depth can effectively improve the strength and the strain force of small piece glass after cutting.

The utility model discloses an adjustable bowl-shaped slotting cutter mounting device, which relates to the technical field of slotting processing, and comprises a taper shank seat, the bottom of the taper shank seat is provided with an inner spigot, the outer wall of the taper shank seat is annularly provided with a plurality of adjusting holes at intervals, and each adjusting hole is communicated with the inner spigot; a slotting cutter is fixedly connected to the bottom of the cutter sleeve, and a boss is arranged at the top of the cutter sleeve and can be inserted into the inner spigot; each fixing screw is used for connecting the taper shank seat and the cutter sleeve; and each adjusting screw is in threaded connection with one adjusting hole. The taper shank seat is mounted on a gear shaping machine through the Morse taper shank and the counterbore, the cutter sleeve and the gear shaping cutter are mounted together, then the cutter sleeve is mounted on the taper shank seat through the square boss and the fixing screws, and the gear shaping cutter can be adjusted by loosening and tightening the adjusting screws at all positions, so that the adjustment efficiency of the bowl-shaped gear shaping cutter is improved, and the adjustment precision of the gear shaping cutter is improved. And meanwhile, the circular boss at the connecting part of the cutter sleeve and the slotting cutter can be freely designed, so that the universality of the slotting cutters with different standards is improved.

The utility model provides a correction device which is used for a sighting telescope system, and the sighting telescope system comprises a white light sighting telescope and a front infrared sighting telescope. The correction device comprises an optical platform, and a light splitting module, a first eye space camera and a second eye space camera which are respectively arranged on the optical platform. The light splitting module comprises a light inlet channel, a first light outlet channel and a second light outlet channel. The white light sighting telescope is arranged on the second light outlet channel, the second eye space camera is arranged at the eyepiece end of the white light sighting telescope, the front infrared sighting telescope is arranged on the light inlet channel, and the first eye space camera is arranged on the first light outlet channel. The first objective camera is used for acquiring and outputting a first image, and the first image is a full-view-field image of the sighting telescope system; the second eye camera is used for acquiring and outputting a second image, and the second image is a divided image of the white light sighting telescope. According to the correcting device, the front infrared sighting telescope can be conveniently adjusted, and the adjusting efficiency is improved.

This utility model relates to an adjustable suspension arm bushing, comprising an inner bushing tube, an outer bushing tube, and a rubber body. The outer surface of the inner bushing tube is threaded, and the outer bushing tube is coaxially sleeved on the inner bushing tube. The rubber body is bonded between the inner and outer bushing tubes, and both ends of the rubber body have axially protruding bosses. A first limiting block and a second limiting block are screwed onto both ends of the inner bushing tube, respectively. By adjusting the screw-in depth of the first and second limiting blocks, the compression of the bosses is changed, thereby adjusting the axial stiffness of the suspension arm bushing. The advantages of this utility model are that it can reduce development costs and time, and improve adjustment efficiency.

The present invention discloses a quantitative evaluation method for AQI contribution rate based on ML and DM, and the method includes data collection, data preprocessing, creating a quantitative evaluation model for air quality index contribution rate, tuning the contribution rate quantitative evaluation model and quantitative evaluation of air quality index contribution rate. The present invention relates to the field of environmental detection technology, and specifically to a quantitative evaluation method for AQI contribution rate based on ML and DM. This scheme introduces a self-attention layer to capture the dependencies between features, improves the accuracy and generalization ability of the model, designs asymptotically stable activation functions, introduces residual connections, designs contribution rate calculation layers and loss functions, accelerates the convergence speed of the model, improves the model's ability to evaluate feature contribution rates, and improves the robustness of the model; by designing an optimized guide function, an optimized amplitude function, and a mixed amplitude optimized search function, the efficiency and effect of hyperparameter tuning are improved, and the model performance is improved.

The utility model discloses a six -sided top press top hammer regulation and correction positioning device relates to superhard material synthetic equipment technical field, the utility model discloses a square main part, the square main part inner chamber is provided with positioning assembly, positioning assembly one side is provided with the dismantling subassembly. The utility model positioning assembly through will battery with mounting block, laser emitter form linkage power supply structure, and laser emitter keeps parallel with square main part edge, so that the regulation and correction top hammer need not closely standard block, and laser emitter can directly to each direction top hammer projection clear light spot, avoid top hammer advance and block sight, realize the synchronous inspection of many top hammer, need not repeatedly start and stop the work cylinder adjustment position, open laser emitter, form two laser light spots on every top hammer, mark position with marking pen, when subsequent inspection, through the relative position of light spot and corresponding mark point, can confirm the deflection direction and amplitude of six top hammer, and regulation and correction efficiency greatly promote.

The utility model discloses a runner supporting structure convenient to disassemble, install and debug, which comprises a runner limiting assembly, an auxiliary installation assembly is arranged on the side part of the runner limiting assembly, an auxiliary limiting assembly is arranged on the auxiliary installation assembly, a runner assembly is arranged on the auxiliary limiting assembly, and the runner assembly is arranged on the runner limiting assembly. A fastening assembly is arranged on the auxiliary mounting assembly, and a rapid adjusting and fixing assembly connected with the runner limiting assembly is arranged on the auxiliary mounting assembly. Compared with the prior art, the flow channel supporting structure has the advantages that the flow channel supporting structure is ingenious in structural design, high in practicability and convenient to mount, dismount and adjust, the front-back position of the flow channel can be rapidly adjusted when the model of a produced lead frame is replaced by applying the flow channel supporting structure, and the machine adjusting efficiency is effectively improved.

The invention discloses a convolution operator shape parameter processing method and device, equipment, a medium and a product, and relates to the technical field of image processing, and the method comprises the steps: determining each convolution operator shape parameter of a convolution operator in a to-be-optimized image processing project; determining an initial parameter value combination of each convolution operator shape parameter, and determining a parameter search space of each convolution operator shape parameter based on the initial parameter value combination; determining a plurality of groups of candidate parameter value combinations of each convolution operator shape parameter based on the respective parameter search space of each convolution operator shape parameter, operating each group of candidate parameter value combinations in the to-be-optimized image processing project, and determining operator execution time consumption of the convolution operators under each group of candidate parameter value combinations; and assigning the candidate parameter value combination with the shortest operator execution time consumption in each group of candidate parameter value combinations to each convolution operator shape parameter. According to the method and the device, the tuning efficiency of tuning the shape parameters of the operator is improved.

This invention discloses a multi-model adaptive fusion method and system for predicting flower price indices, belonging to the fields of smart agriculture and flower big data technology. Addressing the strong seasonality, holiday effects, significant varietal differences, and lag in price transmission characteristic of flowers, this invention proposes specific data preprocessing: distinguishing outliers from holiday fluctuations and supplementing missing values ​​with industry experience; constructing derived features such as demand intensity and supply risk; designing scenario-driven dynamic model routing, configuring models and factor weights differently according to short / medium / long-term periods, with long-term prediction using LSTM+Optuna automatic optimization; automatically matching stable / volatile / small-sample specific models according to flower categories; dynamically weighting and fusing sub-model historical errors, combining regional and channel-specific weights to achieve price conversion across three ends, and constructing an error closed-loop iteration mechanism and engineering optimization. This invention achieves high-precision multi-period prediction with an accuracy ≥95%, adapting to 60+ mainstream flower categories, and supporting full-chain decision-making.

This invention discloses a device for adjusting the arc shape of a vehicle's outward-swinging door and its usage method, belonging to the field of vehicle manufacturing technology. It includes a connecting beam structure and a clamping mechanism. The connecting beam structure has a support beam positioned along the length of the outward-swinging door on its inner side. The clamping mechanism has an L-shaped adjusting guide rod. One end of the adjusting guide rod is slidably mounted on the support beam on the side away from the outward-swinging door, and the other end has a connecting crossbar positioned on the outer arc-shaped surface of the outward-swinging door. The adjusting guide rod and the connecting crossbar form a U-shaped structure, with its open end positioned on the side of the outward-swinging door. The connecting crossbar can move closer to or further away from the outward-swinging door along the adjusting guide rod. A clamping component is mounted on the connecting crossbar, symmetrically arranged with the support beam along the outward-swinging door. This invention enables visual comparison and rapid, precise adjustment of the outward-swinging door's arc shape against the vehicle frame, improving the consistency of product adjustment quality and production efficiency.

The invention discloses an intelligent decision algorithm model design and development auxiliary system, and relates to the technical field of computer software, and the system comprises a data flow and feature engineering pipeline which is used for automatic data processing and feature construction; the model factory and hyper-parameter autonomous optimization engine is used for automatically generating and optimizing a model based on Bayesian optimization; the multi-dimensional evaluation and decision-making cockpit is used for carrying out comprehensive performance evaluation and visual comparison on the candidate models; and the full-life-cycle tracing and knowledge graph module is used for systematically recording and managing experimental knowledge. The invention aims to solve the problems of fragmentization of a model development process, low hyper-parameter tuning efficiency and weak experiment management, realizes full-process automation and intelligent coordination from data to model deployment, improves development efficiency and model performance, and promotes precipitation and reuse of knowledge.

The invention discloses a soil arsenic content spectrum prediction model based on a continuous projection algorithm, and belongs to the technical field of soil arsenic content prediction. A technical route of overlapping peak separation preprocessing, continuous projection algorithm feature screening and multi-source feature fusion to AdaBoost lightweight ensemble learning is adopted, and the technical problems of spectrum overlapping peak interference, complex model parameter adjustment, neglecting of the influence of mining activities and deformation on the arsenic content and the like are solved. According to the method, the accuracy and robustness of soil arsenic content prediction are improved, the model debugging process is simplified, and the method is suitable for soil arsenic pollution accurate monitoring and ecological risk assessment of the grouting and filling mining area.