37results about How to "Reduce precision loss" patented technology

The invention discloses a quantitative training method and system for a deep neural network, and the method comprises the steps: obtaining an original neural network model and a verification data set, and generating a quantitative priority through hierarchical sensitivity evaluation and topology analysis; on the basis of the layer sensitivity matrix, differentiated quantization parameters are configured for each layer, and quantization constraints are established; constructing an error compensation mechanism, and generating an error compensation strategy; executing dual-scale adaptive quantization training, and respectively adopting an independently optimized quantization strategy for forward propagation and back propagation; dynamically adjusting quantization parameters according to the training state and the resource constraint; and evaluating model performance and resource efficiency. According to the method, through multi-dimensional sensitivity evaluation, a dynamic quantification strategy and system error compensation, the precision and efficiency of quantification training are improved, and meanwhile it is guaranteed that deployment resource constraints are met.

The invention discloses a self-adaptive pruning method based on priority experience playback. The method comprises the steps of training a to-be-pruned image classification model, converting a pruning process of a neural network into a Markov decision process, and designing a corresponding state space, an action space and a reward function; constructing and initializing a value network, a strategy network, a target value network and a target strategy network of a depth deterministic strategy gradient algorithm; pruning the to-be-pruned image classification model; storing the pruning process of the to-be-pruned image classification model into an experience playback pool, and setting a priority for the pruning process; updating the sampling probability of each sample according to the time difference error; updating the value network and the strategy network, updating the target value network and the target strategy network, and carrying out loop training until a set maximum number of iterations is reached; and carrying out fine tuning on the image classification model to be pruned of image classification. According to the method, the pruning proportion of each layer of the network can be adaptively generated, and adaptive compression of the network to-be-pruned image classification model is realized.

The invention discloses a rapid marine ship target identification method based on YOLOv8, and is applied to the technical field of target identification. Comprising the following steps of collecting an image sample of a ship navigation water area in a digital vehicle, cleaning and marking the image sample, performing data enhancement on the image sample to obtain a data set which can be used for training and verification, constructing a target identification network model based on a YOLOv8 deep neural network, and obtaining a target identification result by setting hyper-parameters such as a learning rate and a batch processing size. And carrying out transfer learning by using the pre-training model, removing a repeated detection frame by using a non-maximum suppression algorithm and carrying out result filtering by using a confidence coefficient threshold to carry out post-processing on a reasoning result of the model so as to complete prediction of the ship target, and carrying out sparse pruning training on the trained model neural network so as to obtain the ship target. And carrying out calibration and quantification on the pruned model. According to the method, on the premise that the model detection precision is reserved to the maximum, memory consumption during equipment operation is reduced, and meanwhile the model detection speed is increased.

The invention discloses a three-degree-of-freedom macro-micro composite piezoelectric nano probe station, and relates to the technical field of semiconductor manufacturing and biotechnology. In order to solve the technical problems that in the prior art, an existing motor-driven probe station is large in cooperative control difficulty, the positioning precision and the stability are difficult to consider, the structure is complex, and the energy consumption is high, the technical scheme provided by the invention is that the three-degree-of-freedom macro-micro composite piezoelectric nano probe station is provided. Comprising a probe bench surface, an XY macro-motion mechanism, an XY micro-motion mechanism and a Z-direction mechanism. The XY macro-motion mechanism comprises an X piezoelectric motor and a Y piezoelectric motor which are used for achieving dozens of millimeter level horizontal movement in the X-axis direction and the Y-axis direction respectively. The XY micro-motion mechanism is used for realizing accurate movement of several nanometers to dozens of nanometers in the X-axis direction and the Y-axis direction; the Z-direction mechanism is used for realizing dozens of microns of lifting in the Z-axis direction; the probe bench surface is used for bearing a bearing object and is also used for driving the bearing object to move along the X-axis and Y-axis directions. The device is suitable for semiconductor manufacturing and biotechnology testing work.

This utility model discloses a single-station hot plate welding machine, belonging to the field of hot plate welding technology. The hot plate welding machine includes a cabinet, an upper mounting part, a lower mounting part, and a heating component. The movement of the upper and lower mounting parts is driven by first lead screws driven by first servo motors on both sides. The synchronous control of the dual motors can achieve precise synchronization of movement in opposite directions, and the lead screw drive can provide continuous and stable mold closing pressure. Compared with the traditional slide bar drive, it avoids welding gaps caused by pressure fluctuations, ensures tight contact of the workpiece welding surface, and improves welding strength and consistency. The linear slide rail with built-in ball bearings, cage, and return channel is used in conjunction with the sliding block to replace the sliding friction of the traditional slide bar. The low friction coefficient of rolling friction can significantly reduce the wear of the slide rail and the slider, avoiding the problem of precision decay caused by the thinning of the slide bar and the enlargement of the sliding hole. At the same time, the multi-directional contact design of the slide rail can limit the movement of the upper and lower mounting parts, ensuring accurate workpiece alignment during mold closing.

The embodiment of the present invention provides a high-speed and high-precision single-drive gantry structure, including an equipment frame, a first crossbeam assembly and a second crossbeam assembly, wherein the first crossbeam assembly and the second crossbeam assembly are slidably arranged on the equipment frame, and the equipment frame includes a first active side guide rail and a first driven side guide rail, and the first crossbeam assembly is slidably connected to the first active side guide rail and the first driven side guide rail at both ends along the X direction. The high-speed and high-precision single-drive gantry structure provided by the present invention can carry the first crossbeam assembly and the second crossbeam assembly at the same time. Here, the two-layer structure has good casting stability and meets the processing technology requirements for high precision requirements, and greatly reduces the precision loss caused by the impact of the acceleration and deceleration process during high-speed movement.

The invention provides an intelligent super surface wave angle of arrival estimation method and device, and the method comprises the steps: dividing a search space into a plurality of grids, and enabling each grid to correspond to a codebook; traversing all codebooks by using an intelligent metasurface array, and counting power information reported by a receiving end; screening maximum power information in the power information, and determining a preliminary direction of arrival according to the maximum power information; dividing the intelligent metasurface array into a plurality of sub-arrays, and performing phase assignment on each array of the sub-arrays based on the initial direction of arrival; and calculating the direction of arrival according to the assigned phase value of each oscillator. According to the invention, the problem of inaccurate array phase acquisition of the RIS under the condition of weak coverage is effectively solved, and high-precision DOA estimation of the RIS is realized.

The application provides a hardware perception-based model quantization method, system, medium, product and terminal, a hybrid execution mechanism is constructed between a high-level software framework and a bottom layer hardware behavior model, only core calculation operators in a quantization calculation graph of an initial quantization model which need to be executed on target hardware are handed over to a hardware behavior model for bit precision simulation, real calculation characteristics of the target hardware are simulated in a quantization process, and other auxiliary, non-computation-intensive operators are reserved in the high-level software framework for execution, so that the simulation verification efficiency is improved, the simulation precision and the execution efficiency are considered, and the goal that the quantized model runs more efficiently and has lower precision loss on the target hardware is achieved.

This invention provides an approximate 4-bit lookup table multiplier, relating to the field of integrated circuits, comprising: a conversion module, a first-stage multiplication module, and a second-stage multiplication module. During the multiplication process, for two compressed partial products with the same input, a 6-input, 2-output lookup table (LUT6_2) is used for simultaneous calculation. For three compressed partial products with the same input, the precision loss of each compressed partial product is calculated, and a 6-input, 2-output lookup table (LUT6_2) is used to simultaneously calculate the two compressed partial products with the smallest precision loss. The compressed partial product with the largest precision loss is calculated separately using a 6-input lookup table (LUT6). For compressed partial products with an input number exceeding the lookup table's input limit, and for unprocessed partial products, a partial carry-truncation method is used to calculate the product result. This invention can improve multiplier performance while reducing multiplier precision loss.

The invention provides a key value caching method and device, electronic equipment and a storage medium. The method comprises the steps of obtaining a plurality of text segments corresponding to a current text and a segment attention score corresponding to each text segment; based on the fragment attention score, screening at least one target text fragment from the plurality of text fragments; performing attention analysis on the plurality of candidate lexical elements in the at least one target text fragment to obtain first attention scores corresponding to the plurality of candidate lexical elements; based on the first attention score, screening out at least one target lexical element from the plurality of candidate lexical elements; and updating the original cache based on the key value cache corresponding to the at least one target lexical element. By utilizing the embodiment of the invention, the recall of the key value cache of the target lexical element can be realized, the information loss caused by relatively far lexical element information is avoided, the reasoning of relatively long context can be supported, and the problem of precision loss caused by information loss in a rarefaction process is alleviated.

The utility model relates to a kind of mobile press pin machine devices, including product placement unit, cylindrical pin positioning guide unit and mobile press pin machine unit;Wherein product placement unit includes the mobile trolley of placing camshaft product;Cylindrical pin positioning guide unit includes cylindrical pin positioning guide sleeve, locking screw and cylindrical positioning bolt;Mobile press pin machine unit includes product positioning seat, press pin head, gas-liquid pressure cylinder front seat, gas-liquid pressure cylinder, gas-liquid pressure cylinder rear seat, cylindrical guide rail, flange linear ball bearing, gas-liquid pressure cylinder bottom plate, round flange seat, lifting mechanism bottom plate frame, directional wheel, universal wheel, cylinder, bevel gear elevator, hand wheel, handle, pneumatic element bottom plate, gas source two joint parts, solenoid valve and controller.The utility model can improve the precision and efficiency of camshaft workpiece press-fit positioning pin, and the utility model can reduce workpiece collision risk and personnel safety hazard.

The present invention proposes a model generation and real-time construction method for unsupervised learning models, including: step 1, obtaining the original model; step 2, generating a model block, obtaining a derived model block according to the model block, and constructing a training environment for each derived model block; step 3, training the derived model block and updating the parameters of the derived model block; step 4, evaluating all model blocks to quantify the performance indicators and use value of the model block; step 5, modeling the edge reasoning scenario, constrained optimization problem; step 6, solving the constructed optimization problem; step 7, generating a new model according to the solved model block selection scheme, and scaling the model. The method avoids the influence of instability in unsupervised learning on the generation of derived models, ensures the high precision of derived models; reduces the offline generation cost of derived models in unsupervised learning and the online cost of real-time scaling models in edge reasoning.

The utility model discloses a photovoltaic power station support rotating part of automatic lubrication function belongs to photovoltaic power station support technical field, it includes base, the base upper end surface center place is equipped with stand, the stand upper end surface center place is equipped with rotating structure, and the rotating structure includes rotating frame, and the both sides wall center upper of rotating frame is equipped with bearing evenly, and the rotating frame inside center upper is equipped with rotating rod, in addition, the utility model discloses, realize through the linear motion of hydraulic telescopic link efficient stably conversion is driven support main body rotation's circular motion to this adjustment photovoltaic board angle better receive sunlight, has reduced equipment cost, has reduced the precision loss of the intermediate transmission link, just control the telescopic amount of hydraulic telescopic link, can accurate control rotating frame and support main body's rotation angle, guarantee photovoltaic board can continuously track sunlight, effectively improve power generation efficiency.

The present invention provides a lightweight pulse neural network single photon radar back-end signal processing method, including: simulating the optical and electrical processes of a SPAD sensor to collect images, calculating the photon trigger probability of the SPAD sensor; using the photon trigger probability and the grayscale image data set to generate pulse sequence data that obeys the Poisson distribution, and then training the constructed pulse neural network to obtain a trained pulse neural network; quantizing the network parameters of the trained pulse neural network to obtain a lightweight pulse neural network, and then deploying it to a single photon radar signal processing module, and predicting the pulse data collected by the SPAD sensor to obtain classification or imaging results. The present invention can minimize the precision loss of quantization by quantizing the network parameters of the pulse neural network, and converting the network parameters into integers, which can reduce the overhead of computing and hardware resources, and realize high-accuracy, high-efficiency, and high-dynamic range processing of on-chip pulse signals.

This invention discloses a convolutional neural network pruning method and apparatus based on feature similarity and feature compensation. The method includes: adjusting the size of the input images in the dataset to a fixed size, standardizing the pixel values ​​of the images, and increasing the training data using image enhancement techniques; initializing the network structure and setting model parameters, and training the model using the training data; for the trained model, obtaining the similarity between convolutional kernels, performing cluster analysis based on the similarity, grouping each layer of convolutional kernels into similar groups, selecting the convolutional kernels to be retained in each similar group, and generating a new network structure; copying the parameters of the retained convolutional kernels into the new network structure, and compensating the weight parameters of the pruned convolutional kernels in each similar group into the retained convolutional kernels through parameter superposition; performing model accuracy recovery training using the original dataset, and saving the model parameters and network structure. This pruning method can maintain model accuracy.

The present invention discloses an image data processing method, device, equipment and medium. The method comprises: receiving an image to be processed; processing the image to be processed using an image processing convolutional neural network to obtain a processing result of the image to be processed; wherein the image processing convolutional neural network is a network after updating the original convolutional neural network based on each fixed-point tensor in at least one fixed-point tensor and a scaling vector corresponding to each fixed-point tensor; for each fixed-point tensor, the fixed-point tensor and the scaling vector corresponding to the fixed-point tensor are obtained by quantizing each of the multiple first sub-tensors of the weight tensor respectively, and the weight tensor is a tensor of the original convolutional neural network obtained and corresponding to the fixed-point tensor; the multiple first sub-tensors are obtained by decomposing the weight tensor along a predetermined dimension. According to an embodiment of the present invention, it is possible to improve the image data processing accuracy on the basis of having the applicability of the hardware platform.

The utility model relates to metal processing technical field, concretely relates to a kind of multi-station roller needle guide pillar assembly, comprising: sleeve, the inner wall sliding connection of sleeve has roller needle bushing, the inner wall sliding connection of roller needle bushing has column, the outer wall of sleeve is equipped with support device. Through support device, by being equipped with thread hole and limiting bolt in the circumferential, length direction array distribution of sleeve and connecting bracket, the circumferential angle and axial position of support device can be adjusted, cooperate with the relative sliding design of inclination adjusting hole and locking bolt of variable position support, the supporting angle of different stations can be adjusted, through regular polygon roller needle bushing and column, avoid guiding deviation, roller needle bushing converts the sliding friction between column and sleeve into rolling friction, reduce movement resistance and component wear, both meet the requirement of axial positioning in stamping, assembly and other scenes, and it is favorable to reduce precision attenuation after long-term use.

This invention relates to a precision positioning device for roof outer panels with detection and transfer functions, comprising a transfer frame, a positioning frame, a front traction device, two front limiting devices, two auxiliary positioning devices, two front flipping Z-axis protection devices, two rear flipping Z-axis protection devices, two rear limiting and positioning devices, and a rear traction device. This precision positioning device for roof outer panels meets the needs of various functional devices during the production, transportation, and assembly processes of the roof, reducing the number of transfers, improving transfer efficiency, and reducing accuracy loss and deformation risks during the process. The invention features a highly modular structure, allowing for flexible production. It can be used for switching between multiple vehicle models to meet the requirements of the above processes, reducing accuracy loss and sheet metal deformation risks caused by multiple roof transfers. It also provides a direct evaluation of whether the roof surface is qualified and whether it can be normally put on the production line and installed on the vehicle. Roofs for different vehicle models can be flexibly used by switching the overall positioning frame or the support positioning blocks.

The application provides a building missing structure point cloud repairing method and device, the method comprises the following steps: obtaining original point cloud data of a target building; determining missing structure features of the original point cloud data, and adjusting the posture of the target building based on the missing structure features to obtain target point cloud data; constructing a cloth structure corresponding to the target building, determining corresponding cloth node rules based on the cloth structure; performing cloth subsidence simulation on the cloth structure based on the target point cloud data and the cloth node rules to obtain a cloth simulation model; determining structure constraints corresponding to the cloth structure, determining recessed areas in the cloth simulation model that need to be repaired based on the structure constraints, and repairing the recessed areas to obtain regional repair results; and fitting the regional repair results to obtain point cloud repair results of the target building. The application can guarantee strong generalization performance and robust performance of point cloud repair.

The invention discloses a color temperature adjusting method and device for double-primary-color mixed light, equipment and a medium, and relates to the technical field of illumination, and the method comprises the steps: obtaining a first chromaticity coordinate and a second chromaticity coordinate of a first primary-color light source and a second primary-color light source in a specified chromaticity space; obtaining a target chromaticity coordinate of a target color point corresponding to the target color temperature in a specified chromaticity space; vertically projecting the target chromaticity coordinate to a straight line determined by the first chromaticity coordinate and the second chromaticity coordinate to obtain a projection point; determining a first color mixing proportion parameter according to the position of the projection point on the straight line; a second color mixing proportion parameter is obtained based on the projective transformation relation between the color spaces and the edge consistency of projective changes; and determining the duty ratio of the first primary color light source and the second primary color light source according to the second color mixing proportion parameter, the target brightness value and the maximum light-emitting capability of the first and second primary color light sources. The method has good performance on CCT and DUV constraints, and the algorithm calculation complexity is low.

The present invention discloses a pulse neural network quantization method and system for a brain-like computing platform, including: loading and saving a pre-trained full-precision float32 pulse neural network model, recorded as a pulse neural network model T1; modifying the pulse neural network model T1 to obtain a pulse neural network model T2; selecting a calibration data set to train the pulse neural network model T2 in a quantization process, so as to select a suitable scaling quantization factor scale and a zero point zp to minimize the precision loss before and after quantization; deploying the trained pulse neural network model T2 to a brain-like computing platform; and realizing truly asynchronous pulse neural network calculation without precision loss.

The invention discloses a training method of a wireless sensing model, a wireless sensing method and a related device, relates to the technical field of deep learning, and aims to reduce parameter quantity and calculation quantity in a logic gate network training process and reduce training cost based on arithmetic search and calculation of neuron output in a logic gate network. In the process of training the wireless sensing model by adopting the sample channel state information and the tag corresponding to the sample channel state information, the wireless sensing model is trained by adopting a direct-through estimator, so that the precision loss caused by a discretization process on a part of data sets is reduced, and the universality of a logic gate network is enhanced.

The invention discloses a BSD algorithm end-side deployment method based on a YOLO model and capable of running stably. The method comprises the following steps: S1, optimizing a model structure; according to the method, through model structure optimization, three-stage training and export of full-link collaborative design of quantization and end-side deployment, a lightweight and high-precision detection framework BSDYOLO is constructed; mobileNet structured pruning, BiFPN feature fusion and decoupling detection head simplified calculation are adopted, and distillation training, multi-illumination robustness training and quantitative fine tuning are combined to guarantee precision; through a standardized preprocessing process and CPU / NPU collaborative post-processing, stable deployment and real-time reasoning of the model on an end side are realized, the method is suitable for a truck electronic rearview mirror BSD blind area monitoring scene, nine types of targets can be rapidly detected at an edge end with limited computing power, engineering indexes such as detection frequency, response time and missing report rate are met, and a deployment link is stable and reproducible.

The invention discloses a cross-flow flat plate type SOFC single cell surface temperature field prediction method based on tensor CP decomposition, and the method comprises the steps: carrying out the space-time separation of the two-dimensional temperature field tensor of the surface of an SOFC single cell through employing a tensor CP decomposition technology, building a low-dimensional time sequence model of a nominal system, predicting the future low-dimensional feature value, and finally, carrying out the prediction of the temperature field of the surface of the SOFC single cell. Future information of the surface temperature field of the SOFC single cell is predicted in a spatio-temporal data reconstruction mode. According to the method, under the condition that the model prediction precision is guaranteed, the complexity of the prediction model is greatly reduced, the model calculation efficiency is improved, and the problems that a traditional finite element analysis method is long in consumed time and not suitable for online application are solved. The method is particularly suitable for the two-dimensional temperature field modeling problem of the surface of the cross-flow flat plate type SOFC single cell, the temperature field modeling method can better adapt to the modeling and online operation problems of a computing resource limited system, and the application prospect is very wide.

The invention discloses an anti-malicious rapid and safe aggregation method and device, and the method comprises the steps: S1, each of a plurality of participants obtains a plurality of random numbers, and the number of the participants is the same as the number of the random numbers obtained by each participant; s2, each participant generates a corresponding number of pseudo-random numbers for the current aggregation round by using a plurality of random numbers, and calculates the generated pseudo-random numbers to obtain a pseudo-random number calculation result; s3, each participant processes the data to be aggregated according to a pseudo-random number calculation result, and all participants send a processing result to the server for aggregation calculation; and S4, each participant receives the aggregation calculation result returned by the server, and processes the aggregation calculation result to obtain an aggregated result.

The invention provides a main shaft mechanism and a machine tool with the main shaft mechanism, the main shaft mechanism comprises a shaft body assembly, the shaft body assembly comprises a main shaft and a shell, and the shell is internally provided with a cooling channel for circulating cooling liquid; the supporting bearing is arranged on the main shaft in a sleeving mode, a circulation channel is formed in the supporting bearing, and the circulation channel communicates with the cooling channel so that cooling liquid can flow into the circulation channel to cool the supporting bearing; and the sealing component is arranged on the main shaft in a sleeving mode and located on the side of the supporting bearing, at least part of the sealing component is movably arranged in the axial direction of the main shaft, and the problem that in the prior art, the cooling efficiency of the main shaft is low is solved.

The utility model discloses a novel program-controlled multi-tooth indexing table, which relates to the field of program-controlled multi-tooth indexing tables, and comprises an indexing table main body, a motor box is fixedly connected in the indexing table main body, a driving mechanism is fixedly connected in the motor box, one side of the driving mechanism is fixedly connected with a bevel gear A, and the other side of the driving mechanism is fixedly connected with a bevel gear B; the surface of the bevel gear A is engaged with a bevel gear B. The surface of the bevel gear B is fixedly connected with a rotating shaft. Considering that in the prior art, rotation and lifting actions are both completed by an upper fluted disc, so that the system stability and the measurement precision of the multi-tooth indexing table are reduced, a method that lifting and transposition of the fluted disc are completed by a single fluted disc is abandoned, and the lifting action and the rotation action are separately completed by a lifting fluted disc and a rotating fluted disc; the fault rate of the program-controlled multi-tooth indexing table can be reduced, so that the effects of improving the equipment stability and reducing the precision loss caused by a machining process and an assembly process are achieved.

The present invention discloses a sunshine exchange settlement model training method, device and transaction data classification method, which specifically include the following steps: S1: acquiring transaction data and dividing it into a training set and a test set; S2: acquiring the weight of the sunshine exchange settlement model and reshaping it into a second-order matrix; S3: performing full-rank decomposition on the sunshine exchange settlement model to obtain two layers of convolution or reconstruct a fully connected layer; S4: applying orthogonal regularization and funnel regularization on the left singular matrix and the right singular matrix after the full-rank decomposition; S5: adopting an energy-based threshold method to reduce model parameters, thereby obtaining a low-rank model; S6: adjusting the hyperparameters of the low-rank model to obtain a target network model; the target network model is used to classify the transaction data.

The invention provides a video content intelligent analysis method based on an edge computing platform, and the method comprises the steps: collecting a video stream through an image collection interface of the edge computing platform, and carrying out the preprocessing of a video frame at a CPU end, and obtaining standardized image data; the method comprises the following steps: converting a pre-trained neural network model into a special format model adaptive to an NPU (Network Processing Unit), and adopting a hierarchical quantization strategy in the conversion process: quantizing at least one network layer responsible for basic feature extraction in the model by adopting first numerical value precision, quantizing other network layers responsible for classification and regression in the model by adopting second numerical value precision lower than the first numerical value precision; dynamically scheduling the number of processing cores in an activated state in the NPU based on a processing load determined by analyzing the current video frame content represented by the standardized image data, and utilizing the processing cores in the activated state to load a special format model to infer the standardized image data; and post-processing the reasoning output result of the NPU to generate a structured analysis result.