11 results about "Open source technology" patented technology

The invention discloses a model optimization system and method based on deep learning, the system comprises a model square module, a my model module, a model fine tuning module and a model optimization module, and the method comprises the following steps: S1, selecting an initial model; s2, applying an optimization algorithm; s21, a pruning algorithm is applied, and weights or neurons to be removed are determined; s22, a quantification algorithm is applied, and weight and activation are converted into low-precision representation from floating-point numbers; s23, knowledge distillation: selecting a teacher model, defining a student model, and training the student model by using the output of the teacher model as a guidance signal; and S3, selectively performing fine adjustment. Based on a leading edge deep learning method, the open source technology is fully utilized for improvement, and the target of low-cost deployment is achieved under the condition that performance is not lost basically; the model size is reduced, storage is saved, the method is particularly important for mobile equipment or an embedded system, and meanwhile the transmission speed and the deployment speed can be increased.

The invention relates to the technical field of text-to-image generation, and discloses an open source technology Stable Diffusion-based text-to-image generation method, which comprises the following steps of: receiving text or image data input by a user and historical data accumulated by a sugar stacking community, including the historical image data accumulated by the sugar stacking community and image data uploaded by the user (under the condition that the user agrees); the method comprises the following steps: selecting a plurality of data sources and various public information sources, preprocessing input data, including data cleaning, duplicate removal, missing value processing, format conversion, standardization, normalization and other preprocessing steps, normalizing and standardizing numerical values to ensure the quality and consistency of the data, dividing the preprocessed data into a training set and a test set, and algorithm model training is carried out based on a Stable Diffusion model, including deep learning model training, evaluation of the trained model by using a test set, and optimization of the model according to an evaluation result, possibly including hyperparameter adjustment, regularization increase and other strategies.

The invention discloses a GIS-based seismic acquisition basic data comprehensive query platform construction method and device, electronic equipment and a medium. The method comprises the following steps: constructing a seismic acquisition basic database; constructing a B / S architecture platform with separated front and rear ends; gIS map integration is carried out, and then work area boundary information is displayed; integrating retrieval and query functions in the seismic acquisition basic database; and the client requests the data service through the browser. The earthquake acquisition basic data comprehensive query platform is constructed through the open source GIS technology and the database technology, the work area condition in the area can be rapidly analyzed through the platform, the work area related data can be rapidly queried and retrieved, the efficiency and quality of earthquake acquisition design work are remarkably improved, and the method is suitable for large-scale popularization and application. And powerful support is provided for successful implementation of a seismic exploration project.

The present invention discloses an intelligent storyboard script generation system and method, which belongs to the technical field of software engineering, including a front-end interactive interface layer, a communication layer one, a back-end business logic layer, a communication layer two and a computing layer connected in sequence, wherein the front-end interactive interface layer includes text editing and image preview functions, the back-end business logic layer includes a text generation module composed of a large language model (LLM) and an image generation module composed of a text-to-image generation class model (Stable-diffusion), the text generation module and the image generation module process input content to obtain a computationally intensive task, and the computing layer is set as a cluster composed of several servers to calculate the computationally intensive task. The present invention constructs a storyboard script generation system with coherent narrative capabilities and support for open source technology, a text generation function that meets the user's storyboard text creation needs, and can generate images.

The invention discloses a database analysis method based on an open source technology, and relates to the technical field of database analysis, and the database analysis method comprises the following steps: S01, constructing data relation mapping of a target database, and classifying data; step S02, acquiring an internal safety index; s03, acquiring an external security index of the database; step S04, based on the internal security index and the external security index, acquiring a database overall security index, and corresponding to a corresponding security level; and S05, taking different levels of security protection measures based on different security levels. According to the database analysis method based on the open source technology, the database analysis technology based on the open source technology is provided, the database safety is analyzed in detail, a visual and digital basis is provided for safety judgment of the database, safety protection of the database can be enhanced in time, and the safety of the database is improved. And the security of various data in the database is improved.

The invention discloses an intelligent education knowledge graph construction method based on a large language model, and the method comprises the steps: S1, obtaining curriculum text data from an intelligent education video platform, and cleaning and integrating the curriculum text data into a long text document; s2, performing semantic slicing on the long text document by using a natural language processing model to obtain semantic fragments; s3, constructing a course category included in the course knowledge graph, and obtaining extended noun explanation of the course category to obtain an explanation text; s4, finding a corresponding course category for each semantic segment by calculating the similarity between the embedded vectors; s5, decomposing the semantic fragment into a plurality of knowledge points by utilizing a large language model and CoT cue words, and marking; and S6, storing the knowledge points in the neo4j graph database in a defined format to form a multi-course knowledge graph. According to the method, illusion problems and errors caused by completely depending on a large model can be avoided as much as possible, and construction of the course atlas is realized by efficiently utilizing partial open source technical means.

The invention discloses a system and a method for making high-quality large model training data, the system comprises a data set management module, a data annotation module, a sensitive word library module, a data processing module and a data analysis module, and the method comprises the following steps: S1, defining metadata; s2, importing data; s3, converting a data format; s4, storing the data; s5, processing the data; and S6, analyzing the data. Based on a leading edge deep learning method, an open source technology is fully utilized to formulate standards and specifications of training data of a large model, and the data specification problem is solved; a data pipeline for data collection, filtering, arrangement and storage is established, and the data processing efficiency is improved; and data quality evaluation is perfected to collect enterprise core data.

The embodiment of the invention provides a construction CAD webpage loading lightweight method and device and a storage medium, and the method comprises the steps: extracting vector data of a CAD file based on an open source software secondary development API, and exporting the vector data as a general vector format file; performing automatic lightweight processing and format conversion on the general vector format file according to service requirements; and loading the converted lightweight data format file through a front-end open source library to realize webpage loading and interaction. A full-open-source technical scheme is adopted, commercial software authorization is not needed, and the development and use cost is reduced. And all data processing is completed in the local server, so that the data security and autonomy are ensured. Through topological optimization and lightweight conversion of the CAD data, the webpage loading speed and rendering efficiency are improved, common hardware equipment can also run smoothly, and the problems of high cost, low speed and poor safety in the prior art are solved.

The invention discloses a fine tuning system and method based on a deep learning model, and the system comprises a data set management module, a model square module, a my model module, and a model fine tuning module, and the method comprises the steps: S1, managing a basic large model; s2, fine adjustment is carried out on the basic large model; and S3, releasing the new model. Based on a leading edge deep learning method, an open source technology is fully utilized for improvement, high calculation and storage cost is reduced through a modularization and parameter efficient fine tuning technology, data fitting risks are reduced by using a method of fine tuning of a small number of parameters, field training data are constructed, and a large model is generally used on the basis of an advanced. And training a large model with domain expert ability.

This solution relates to a method for the adaptive style of in-vehicle electronic navigation maps based on open-source GIS technology, which solves the problems of inconvenient user use and high development costs existing in the existing map style adaptation solutions. The method includes: running map server software on the background server to publish WMS, and associating WMS with dynamic SLD services and dynamic SVG services through file references; subscribing to the corresponding WMS on the in-vehicle device, and sending user information input by the user to the background server; the in-vehicle device records the scene characteristics and behavior characteristics of the user using the map, and sends them to the background server; the background server inputs the user characteristics including user information, scene characteristics and behavior characteristics of the user using the map into the recommendation system module, so that the recommendation system module maps the user characteristics into map style parameters of SLD files and SVG files, and further enables the dynamic SLD service and / or the dynamic SVG service to generate style files suitable for the user according to the map style parameters; the map is transformed into the corresponding map style.

The application discloses a general artificial intelligence knowledge infrastructure and a multi-modal knowledge processing method, and belongs to the field of artificial intelligence software systems. In view of the problems of existing multi-modal data semantic fragmentation, fragile knowledge storage format and static and rigid knowledge graph, the application provides a full-stack solution through a four-layer software architecture: a virtual unified knowledge bus (V-UKB) realizes automatic discovery and millisecond to microsecond time synchronization of multi-modal devices; a multi-modal knowledge record format (MKRF) realizes atomic-level fusion of perception data and semantic knowledge, and forms self-description, cross-platform and traceable knowledge units; and a Qian Kun knowledge core realizes neural-symbol hybrid reasoning and dynamic evolution of knowledge. All technical solutions of the application are realized based on general computing devices and open source technologies, do not require special hardware support, and can be widely applied to scenes such as intelligent libraries, long-term archives, cross-modal intelligent retrieval, distributed knowledge services and the like.