79 results about "Vector element" patented technology

The invention discloses a natural resource big data off-line comprehensive analysis, annotation and proof system and method, and belongs to the technical field of natural resource informatization and geographic information processing. The system comprises the following components: a spatial data fusion module, which is used for realizing unified management and efficient processing of heterogeneous spatial data; the spatial index optimization module is used for realizing quick retrieval and high-performance query optimization of spatial data; the intelligent annotation and arrangement module is used for realizing low-code rapid configuration and business process automatic arrangement of different natural resource investigation scenes; the off-line analysis module is used for realizing buffer area analysis, space superposition and terrain analysis in an off-line environment; the real-time visualization module is used for realizing multi-level map display, track recording, vector element editing and multimedia evidence collection; and the data quality management module is used for realizing automatic quality inspection of geometry, topology and attributes and statistical analysis, version control and export management of annotation results.

The invention discloses a three-dimensional visualization method, system and device based on heterogeneous spatial data fusion and dynamic sectioning, and a medium, belongs to the technical field of urban information model visualization, and aims to solve the technical problem of how to improve data fusion efficiency, sectioning response delay and mobile terminal frame rate in a million patch scene. Light-weight and high-interaction visual display of scenes such as smart cities and underground pipe galleries is realized, and the defects of multi-source spatial data splitting, high sectioning interaction delay and poor cross-terminal compatibility are overcome. According to the technical scheme, heterogeneous data dynamic normalization is carried out, wherein local coordinates of a building information model, WGS84 coordinates of a geographic information system and computer aided design data are unified to an Euclidean coordinate system through a space dynamic reference conversion algorithm, and normalized data are obtained; constructing a semantic data relation graph; generating a dynamic sectioning plane in real time; vector elements are highlighted in a linkage manner; and carrying out cross-terminal adaptive rendering.

The invention discloses a PDF document-oriented chart area accurate positioning and extraction method, and belongs to the technical field of information processing.The PDF document-oriented chart area accurate positioning and extraction method comprises the following specific steps of 1, analyzing a PDF vector instruction, analyzing a binary stream of a PDF document, extracting a bottom layer drawing instruction set, and extracting a bottom layer drawing instruction set; a set of structured vector elements is generated, each element containing a type, coordinates, and operator attributes. According to the method, through a vector-vision multi-mode fusion mechanism, the chart positioning precision is greatly improved, the error of an extraction region is remarkably reduced through sub-pixel-level boundary optimization, the PDF single page is converted into a vector element set and a visual feature map, and the vector element set and the visual feature map are processed together, so that the single page processing time is shortened, a cross-page chart splicing mechanism supports continuous document processing, and the processing efficiency is improved. And complex formats such as image-text surrounding and semitransparent covering can be effectively processed, the output structured data can be directly used for a data analysis assembly line, and the manual intervention cost is greatly reduced.

A computing system including a hardware accelerator configured to receive vector processing instructions from a processor. The vector processing instructions include an initial read address, an input increment size, and a vector processing operation. During vector processing iterations performed at vector processor tiles included in a vector processor tile array, the hardware accelerator reads vector elements into respective vector processor tiles in an input stream. The vector elements are read into the vector processor tiles from locations in memory that start at the initial read address and advance by the input increment size at successive vector processing iterations. At each of the vector processor tiles, the hardware accelerator computes a vector processing result at least in part by performing the vector processing operation on the vector element read into the vector processor tile. The hardware accelerator outputs the vector processing results from the vector processor tiles in an output stream.

An on-chip memory device within a hardware attention unit of a neural network accelerator, comprising: a) at least one dedicated on-chip memory block (1) for storing key vectors and at least one dedicated on-chip memory block (2) for storing value vectors, the memory blocks being located within the attention unit; b) a validity bitmap (3) with one bit per token position indicating whether valid key / value vectors are stored for the respective position; c) a store logic that, when calculating new token positions, stores the calculated key / value vectors element-wise in the on-chip memory and, after completion of all dimensions, sets the corresponding bit in the validity bitmap (3); d) a load logic that, in subsequent calculations, makes stored key / value vectors available from the on-chip memory with a latency of one clock cycle;unde) a hit check logic (10) which combinatorially (without clock cycle latency) indicates, based on the validity bitmap (3), whether valid K / V vectors are available for a queried token position.;

The invention relates to the field of data lookup, and discloses a locality perception lookup table lookup method, which comprises the following steps of: S1, reordering between rows: clustering and reordering the rows of an input vector and a weight matrix according to the same quantized value to enable elements with the same quantized value to be continuously distributed; for a common matrix GeMV scene, cache locality is improved through rearrangement and subsequent blocking operation, and repeated loading of a lookup table LUT is reduced. And the data locality of the GeMV scene is remarkably improved through inter-line reordering and a virtual mapping strategy. On the basis of reordering of input vector element values, elements with the same value are aggregated, so that the LUT line loading times are reduced to 256 times at most from being in direct proportion to the input dimension d, and frequent replacement of the LUT lines in the WRAM is avoided; meanwhile, a weight matrix is divided into BR * BC sub-matrixes, the characteristic that vertical adjacent rows share an accumulator is utilized, multi-row sub-matrix calculation is loaded at a time, and virtual rearrangement maintains the row matching relation between an input vector and the weight matrix through index mapping.

Disclosed embodiments provide techniques for improved performance in processing vector instructions. A processor core is accessed. The processor core is coupled to a memory hierarchy, and the processor core includes one or more vector execution units (VUs), and one or more load store units (LSUs). The processor core includes a vector register file (VRF). The VRF includes multiple vector registers, and each vector register includes multiple vector elements. Vector elements that have a source or destination in contiguous memory are identified. Load store units (LSUs) take advantage of the contiguous memory condition by executing a vector load or vector store operation as a single memory access, requiring a reduced number of clock cycles. The single memory access satisfies each memory operation for each vector element within the vector register file.

A computing device, including a hardware accelerator configured to train a machine learning model by computing a first product matrix including a plurality of first dot products. Computing the first product matrix may include receiving a first matrix including a plurality of first vectors and a second matrix including a plurality of second vectors. Each first vector may include a first shared exponent and a plurality of first vector elements. Each second vector may include a second shared exponent and a plurality of second vector elements. For each first vector, computing the first product matrix may further include computing the first dot product of the first vector and a second vector. The first dot product may include a first dot product exponent, a first dot product sign, and a first dot product mantissa. Training the first machine learning model may further include storing the first product matrix in memory.

The application discloses a raw oil property prediction and analysis method, device, equipment and a storage medium, and comprises the following steps: generating a raw oil standard comparison model in the form of a vector matrix according to a property data sample set comprising property data of a large number of crude oil samples; each sample vector group in the vector matrix comprises vector elements including distillation point, equivalent double bond value and property data; when performing property analysis and prediction on the crude oil to be analyzed, a vector group comprising vector elements of distillation point and equivalent double bond value is generated according to the crude distillation point temperature data in the simple evaluation report and the carbon, hydrogen and nitrogen element content and the mixture molecular mass data under the corresponding distillation point; then, the corresponding similar sample is determined through the raw oil standard comparison model by means of Euclidean distance calculation, and the property data of the crude oil to be analyzed is predicted; and the application can greatly improve the detection and analysis efficiency, thereby improving the timeliness of oil product detection.

The application provides a geometric figure processing method and device, equipment and a storage medium. The method comprises the following steps: acquiring a to-be-recognized image, and recognizing and cutting the to-be-recognized image to determine an initial geometric illustration; performing optimization processing on the initial geometric illustration to determine a target geometric illustration; performing denoising processing on the target geometric illustration to obtain a denoised geometric illustration, acquiring image features of the denoised geometric illustration, classifying the image features to determine a target geometric element, and determining a first vector element according to the target geometric element and a geometric element relationship obtained by disassembling the target geometric element; editing a to-be-edited vector element in the first vector element based on an editing instruction, determining a vector figure according to the to-be-edited vector element and a non-editing vector element, and exporting the vector figure to a figure resource library in a standard format. Thus, high-precision recognition is performed on the geometric figure, and the subsequent labor cost for adjusting the geometric figure is greatly reduced.

The application discloses a vector tile-oriented cascade generation and dynamic optimization method and device, relates to the technical field of geographic information data processing, and can solve the problem of how to simultaneously improve generation efficiency, guarantee mapping synthesis quality and control tile data volume in vector tile multi-scale production. The scheme comprises the following steps: acquiring input element data; the input element data of the highest precision level is original vector data, and the input element data of the remaining levels is the vector tile generated in the last level; constructing synthesis priority based on multiple dimension characteristics of elements for the input element data, and screening the input element data according to the priority; dividing and encoding the screened elements according to the tile grid of the current level to generate temporary vector tiles; when the temporary vector tiles exceed the preset data volume constraint, converting the vector elements in the temporary vector tiles into a raster expression and performing iterative thinning until the constraint is met, and then outputting the temporary vector tiles as the target vector tiles of the current level.

The application relates to a logic arrangement prediction model training method, a logic arrangement method and device, computer equipment, a storage medium and a computer program product, and relates to the technical field of computers and can be used in the field of financial technology and other related fields. The method comprises the following steps: constructing a syntax tree of a sample logic for each sample logic in a low-code platform; determining a logic sequence vector of the sample logic based on the syntax tree; performing sliding window processing on the logic sequence vector to obtain at least one logic sequence subvector of the sample logic; confirming a component corresponding to a next vector element of the logic sequence subvector in the logic sequence vector as a target component of the logic sequence subvector; and training a logic arrangement prediction model to be trained by taking the logic sequence subvector as input information and taking the target component of the logic sequence subvector as supervision information, so as to obtain a trained logic arrangement prediction model. The method can improve the logic arrangement efficiency in the low-code platform.

This application proposes a method, apparatus, electronic device, and storage medium for arranging vector elements. The method includes: acquiring vector elements to be processed, and determining the levels of each switching node (SN) unit in a butterfly network based on the number of vector elements to generate the butterfly network; wherein the butterfly network is used to arrange the vector elements; determining the input-output mapping relationship between the levels of each switching node (SN) unit in the butterfly network; determining the target state of the SN unit in each level based on the input-output mapping relationship; arranging the vector elements based on the target state, and performing data processing based on the arranged vector elements. Therefore, this scheme enables the arranged vector elements to be transmitted non-blockingly in the butterfly network, thereby enabling the mapping of complex and variable large model workloads onto vector hardware to achieve efficient data processing execution.

The invention discloses a method and device for recognizing and reconstructing vector characters of a CAD drawing, and relates to the technical field of data processing. The method comprises the steps that a PDF file is analyzed, and vector elements and attribute information of the vector elements are extracted; character stroke candidate lines are screened out according to attribute grouping; carrying out connectivity analysis and density-based spatial clustering on the lines, and identifying a text region; rendering the vector lines of the text region into an image with high resolution, carrying out single character segmentation, and recognizing characters by combining drawing-level font library template matching with a single character recognition model; combining the single characters into word groups according to the positions and the context information, and performing word order correction by adopting a language model; the recognition result is written back into an editable CAD text object; according to the method, vector geometric information is directly utilized, distortion of bitmap sampling is avoided, the small character recognition rate and the anti-jamming capability are effectively improved, and end-to-end semantic recovery from vector lines to searchable and editable texts is achieved.

The invention provides a vector geographic data efficient visualization method and system based on multi-scale optimization and containerization architecture, and belongs to the field of spatial big data visualization and cloud native GIS application. According to the method, the filling rate of the enclosing rectangles and the number of the vertexes are calculated for the original geographic element data, so that automatic classification of regular elements and irregular elements is realized; and respectively constructing two sets of multi-scale expression systems, and storing four types of multi-scale data tables in a spatial database. A containerized micro-service architecture is adopted, and dependence management and cooperative operation are automatically completed through a container arrangement tool. The vector slicing service dynamically selects a corresponding scale expression according to the scaling level. Through cooperative work of a database trigger, a monitoring service and a cache service, the system can calculate an affected slice number range and execute accurate cache deletion when element attributes are updated, and real-time and on-demand slice updating is realized. The method has the beneficial effects that the visualization performance and updating efficiency of massive vector elements are improved.

The invention relates to the technical field of production management, and provides an industrial intelligent production scheduling method and system based on multi-modal data. The method comprises the steps of obtaining multi-mode production data through an intelligent sensor, and extracting feature vectors corresponding to all modes from the production data through a preset neural network; performing semantic alignment and merging processing on the feature vectors through modal parameters among all modals to generate joint vectors; constructing a dynamic constraint network based on vector elements in the joint vector, and updating a node and event relationship in the dynamic constraint network according to real-time event information corresponding to the vector elements; generating a production function based on the dynamic constraint network, and solving the production function to determine production scheduling control parameters; and sending the production scheduling control parameters to a control terminal. Effective integration and utilization of production information are realized, production scheduling is more practical, and production scheduling accuracy and production efficiency are improved.

The invention discloses a PDF document Logo automatic identification method fusing open vocabulary detection and semantic segmentation, and belongs to the technical field of image processing and computer vision. The method comprises the following steps: firstly, performing page rendering and image preprocessing on an input PDF document to extract bitmap information and vector elements; secondly, extracting a brand name, an image note or a semantic tag in the page by using a text recognition and image tag generation model, and generating a text prompt for visual detection; then, inputting the rendered page image and the text prompt into an open vocabulary detection model, and generating a corresponding candidate target box; then, taking the target frame as a prompt to be input into a segmentation model, so as to obtain a pixel-level segmentation mask corresponding to the Logo; and finally, performing post-processing and vectorization operation on the segmentation mask, and outputting a high-precision Logo segmentation result. According to the method, automatic identification and accurate segmentation of the Logo in the PDF document are realized by utilizing a collaborative mechanism of open vocabulary detection and prompt segmentation.

Tensor processing includes obtaining an input tensor, the input tensor including a dimension of index size n, determining an element count, c, based on an indicator, the indicator specified by the instruction, and the element count specifying a number of vector elements on which to perform an artificial intelligence function, obtaining an input vector, of the input tensor, of size n, and performing the artificial intelligence function, the performing the artificial intelligence function including performing the artificial intelligence function on a first c number of elements of the input vector to provide a corresponding c number of elements of an output vector of index size n of an output tensor.

This application provides a storyboard script generation method. The method includes obtaining a reference image used to generate a storyboard script; obtaining an element of the reference image and element vector information of the element; obtaining a vector element matching the element; and generating the storyboard script by setting the vector element on a specified canvas based on the element vector information. According to the storyboard script generation method provided in this application, the element in the reference image is identified, and the matched vector element is correspondingly set on the specified canvas, to obtain the storyboard script in a form of a vector graph. Therefore, it is more efficient and easier to produce the storyboard script, and user experience is effectively improved.

The invention provides a power utilization strategy recommendation method and device, and relates to the technical field of power utilization load optimization and adjustment. The method comprises the steps of firstly obtaining a plurality of historical power consumption data queues; then extracting data from the plurality of historical power consumption data queues for multiple times, constructing a first historical vector for the data extracted each time, and performing dimension raising conversion and vector element weighting processing on each first historical vector to obtain a plurality of second historical vectors; secondly, for each second historical vector, respectively calculating the similarity with a plurality of class centers, and adding the plurality of similarities and a plurality of electricity consumption predicted values determined by the typical electricity consumption into a predicted electricity consumption data queue; and finally, recommending a power utilization strategy to the user according to the target power utilization number queue and the predicted power utilization data queue. According to the embodiment of the invention, the power consumption curve is predicted based on the double-cycle characteristic of the power consumption curve, and the power consumption strategy is recommended to the target user, so that a better power grid load optimization effect can be realized.

The invention relates to a NumPy numerical calculation optimization method and system based on RISC-V vector extension and application, belongs to the technical field of computer science and high-performance calculation application, and is used for performing high-performance optimization on a core calculation part which does not call a BLAS library, such as vector element-by-element calculation and matrix element-by-element calculation. On the basis of analysis and calculation steps, the performance of NumPy in broadcast operation and element-by-element operation scenes is improved through technologies such as underlying code substitution, loop optimization, data loading and memory alignment optimization, so that support is provided for scientific calculation and high-performance calculation based on the RISC-V architecture.

The invention relates to the technical field of application program front ends, in particular to a 3D map multi-layer dynamic interaction and vector element control simulation method, device and equipment and a medium, and the method comprises the steps: receiving longitude and latitude data required by map vector elements, and configuring a setOption method of ECharts based on the received data; monitoring an interaction event of a map geographic coordinate system by utilizing a response type mechanism of Vue, and establishing a dynamic association relationship; according to the change of the scaling value, dynamically adjusting the configuration of the geographic coordinate system and the data series, and synchronously updating the position of each layer based on the association relationship; traversing vector element data in the data series, and calculating visual parameters of vector elements based on the scaling values; correcting the coordinates of the vector points in real time by adopting a boundary constraint algorithm; and automatically adjusting the icon size and the font size according to the zoom value, and hiding the secondary elements when the zoom level is lower than a threshold value. The position constraint of vector elements is realized, and the visual effect is improved.

Mechanisms are provided for compressing ciphertext data for data transmission. A sparse vector is received, comprising a plurality of vector elements and a tree is built from the sparse vector where each leaf node corresponds to a vector element in the sparse vector, and each subsequent level of the tree is built from a child level below it in the tree. Nodes of a subsequent level have values determined based on values of child nodes connected to them. The mechanisms execute a level-based copy-and-recurse operation on the tree from a root node of the tree to leaf nodes of the leaf node level. The level-based copy-and-recurse operation computes, at each level of the tree, an indicator vector and a selection matrix that identifies which nodes to recurse into. The mechanisms generate the compressed ciphertext data based on the indicator vectors and the sparse vector.

The application provides an optical computing device, method, equipment and storage medium. The device comprises a processor configured to obtain a target vector and a target matrix to be subjected to multiplication calculation; the processor is further configured to decompose the target matrix to obtain at least one sub-matrix and a weight vector; a modulator configured to configure a phase difference of an optical computing unit in a corresponding optical computing array based on an element attribute of a matrix element in the sub-matrix; a laser configured to generate an optical signal; the modulator is further configured to configure a phase of the optical computing unit in the corresponding optical computing array based on an element attribute of a vector element in the target vector; an optical computing array configured to receive the optical signal and output a corresponding optical computing result; and the processor is further configured to perform weighted calculation on the optical computing result based on the weight vector to obtain a calculation result vector of the target vector and the target matrix. The application can reduce the complexity of the configuration of the optical computing unit and avoid information loss.

This invention discloses a deep learning-based intelligent architectural floor plan parsing system, relating to drawing recognition technology. This deep learning-based intelligent architectural floor plan parsing system includes: a data acquisition and preprocessing module for acquiring, converting, and enhancing architectural floor plans from multiple sources; a multi-scale feature extraction module for parallel multi-scale extraction of architectural segmented image sets; a multi-scale feature fusion module for performing scale anomaly compensation processing and feature fusion path selection; an object detection and classification module for detecting reliability through fused features, probability matrices, and cross-scale detail discrimination; a spatial relationship reasoning module for constructing candidate topologies based on vector elements and constructing topology consistency discrimination through relationship reasoning; and a result post-processing and optimization module for achieving cross-segment alignment and topology cleaning. This system solves the problems of unstable extraction and alignment of multi-scale features in architectural floor plans, leading to missed detection of small components and unreliable spatial relationships.

The invention discloses an identity determination method and device, a storage medium and an electronic device, and relates to the technical field of smart home, and the identity determination method comprises the steps: obtaining echo signals collected by N radars located at a target position, and determining a three-dimensional signal matrix according to the N echo signals, N being an integer greater than or equal to 2, the dimensions of the three-dimensional signal matrix comprise a distance dimension, an angle dimension and a Doppler frequency dimension; a target feature vector of a target object located at the target position is determined according to the three-dimensional signal matrix, and vector elements in the target feature vector comprise the radar scattering sectional area, the gravity center distance, the gravity center angle, the breathing frequency, the heart rate frequency and the phase difference between the N radars of the target object; and based on a preset feature vector library, determining a target identity identifier of the target object according to the target feature vector of the target object. By adopting the technical scheme, the problem that the identity of the target object cannot be accurately identified is solved.

The present disclosure relates to decoupling atomicity from operation size. In one embodiment, a processor implements an atomicity size that is different from an operation size (for memory consistency order). More specifically, the processor can implement an atomicity size that is smaller than the operation size. For example, for a multiple register load, the atomicity size can be a register size. As another example, a vector element size can be the atomicity size for a vector load instruction. As yet another example, a number of contiguous vector elements (but less than all vector elements in a vector register) can be the atomicity size for a vector load instruction.

A computing device, including a hardware accelerator configured to train a machine learning model by computing a first product matrix including a plurality of first dot products. Computing the first product matrix may include receiving a first matrix including a plurality of first vectors and a second matrix including a plurality of second vectors. Each first vector may include a first shared exponent and a plurality of first vector elements. Each second vector may include a second shared exponent and a plurality of second vector elements. For each first vector, computing the first product matrix may further include computing the first dot product of the first vector and a second vector. The first dot product may include a first dot product exponent, a first dot product sign, and a first dot product mantissa. Training the first machine learning model may further include storing the first product matrix in memory.

An apparatus comprises instruction decoding circuitry and processing circuitry. In response to a permute prefix instruction specifying at least one permute source vector register and a destination vector register, the instruction decoding circuitry is configured to generate a micro-operation to control the processing circuitry to perform at least a permute operation to permute vector elements of at least one permute source vector operand specified by the at least one permute source vector register to generate a permuted vector operand corresponding to the destination vector register. The permute prefix instruction has an encoding providing an instruction fusion hint indicating to the instruction decoding circuitry that a next instruction in program order after the permute prefix instruction is expected to be a vector processing instruction specifying, as a source vector register for defining a source vector operand to which a vector processing operation is to be applied, the destination vector register of the permute prefix instruction.