150results about How to "Guaranteed independence" patented technology

The invention discloses a data isolation and privacy protection method and system for a data security large model, and relates to the technical field of artificial intelligence, the method comprises the following steps: setting an independent storage area for each enterprise in an external knowledge base in advance, and transmitting encrypted data to the corresponding storage area based on an enterprise identifier, setting corresponding access authority for the encrypted data of each storage area; judging whether the user identity information and the user access permission information are compliant or not, extracting a keyword corresponding to the user demand information under the compliant condition, and performing similarity retrieval in an external knowledge base based on the keyword to obtain corpus data with the highest similarity with the keyword; the corpus data and the user requirements are jointly input into the large data security model to generate the data security document, and the problems that in the prior art, data security isolation and privacy protection cannot be achieved, and security and compliance use of data in the whole process cannot be guaranteed are solved.

The invention discloses a multi-person collaborative programming method and system, electronic equipment and a medium, and the method comprises the steps: providing a canvas task designer which supports the visual design of a plurality of canvases, and enabling each canvas to correspond to a subtask; sub-tasks designed in one or more canvases are stored as independent module files, and each module file can be developed by different developers; persistently storing the developed module file locally, and creating a module ID for the module file; and importing each module file into a unified project, and editing a data reference relationship and execution logic among the module files. According to the multi-person collaborative programming method and system, the electronic equipment and the medium provided by the invention, the whole project is divided into a plurality of independent sub-tasks through the canvas task designer, and each sub-task is stored in the form of the module file and can be independently developed by different developers.

The invention belongs to the technical field of computer vision and artificial intelligence, and discloses a combined zero sample image classification method based on hierarchical feature fusion. According to the method, a solution based on hierarchical feature fusion is innovatively provided for the technical difficulty of combined zero sample image classification. According to the method, features of different depths in a CLIP model visual encoder are selected and subjected to hierarchical fusion, multi-level and multi-scale visual information is effectively extracted, the fused features are further used for visual and semantic cross-modal interaction, and the understanding ability of the model for complex combinations is remarkably improved. In addition, by introducing a specific loss item, attributes and object features are effectively separated, and the feature aliasing problem is avoided. In a combined zero sample image classification task, the method shows excellent generalization ability and recognition effect, and can significantly improve the classification performance of the model for unknown complex combined images.

The invention relates to the technical field of circuit breaker service life evaluation, in particular to a high-voltage circuit breaker residual service life intelligent evaluation system, which comprises an operation data acquisition module for acquiring operation data of a high-voltage circuit breaker; the degradation degree feature extraction module is used for extracting degradation degree features from the processed operation data; the residual life calculation module is used for constructing a residual life prediction model by utilizing the degradation degree characteristics and adopting a generalized regression neural network algorithm, and calculating the residual life of the residual life prediction model; the early warning and maintenance strategy module is used for formulating a predictive maintenance strategy of the high-voltage circuit breaker based on the residual life calculation result in combination with the residual life probability density function and the expected loss rate of the circuit breaker, and balancing the economical efficiency and the reliability; according to the method, the degradation process of the circuit breaker is reflected in real time, potential faults of equipment are predicted in time, and delay and inaccuracy caused by traditional experience judgment and regular inspection are avoided.

The invention relates to the technical field of photovoltaic cells, and discloses a back contact solar cell and a preparation method thereof. The preparation method comprises the following steps: providing a substrate layer of a first conductive type having a first surface and a second surface opposite to each other; a first groove is formed in the first surface, and a bulge is formed beside the first groove; performing doping treatment on the inner wall surface of the first groove to form a first doping layer; the first doped layer is of a second conductive type; forming a passivation layer on the surface of the first doping layer; a second groove isolated from the first groove is formed in the protrusion; forming a first tunneling oxide layer on the second groove; forming a second doping layer of the first conduction type in the second groove and on the first tunneling oxide layer; a first electrode and a second electrode are formed, the first electrode penetrates through the passivation layer to be connected with the first doping layer, and the second electrode is connected with the second doping layer. The first groove and the second groove which are relatively isolated are formed by grooving at different positions of the substrate layer, so that efficient current collection is facilitated, short circuit is prevented, and the photoelectric conversion efficiency and the power generation efficiency are improved.

The embodiment of the invention discloses an array substrate and a display panel, the array substrate comprises a substrate and a plurality of conductive layers, the plurality of conductive layers at least comprise a first conductive layer, a second conductive layer and a third conductive layer which are laminated, the first conductive layer comprises a grid electrode of a driving transistor and a first polar plate connected with the grid electrode of the driving transistor, and the second conductive layer comprises a second polar plate connected with the grid electrode of the driving transistor. The second conducting layer comprises a second polar plate, the third conducting layer comprises a third polar plate, and the vertical projection of the first polar plate on the substrate and the vertical projection of the second polar plate on the substrate are overlapped and respectively form an upper polar plate and a lower polar plate of the storage capacitor; and the vertical projection of the second polar plate on the substrate and the vertical projection of the third polar plate on the substrate are overlapped and respectively form an upper polar plate and a lower polar plate of the voltage-stabilizing capacitor. According to the technical scheme provided by the embodiment, through the storage capacitor and the voltage stabilizing capacitor which are stacked up and down, the layout space utilization rate is improved, the capacitance value of the storage capacitor is increased, and the display effect is improved.

The invention provides forming equipment for U-shaped groove plates on the two sides. The forming equipment comprises a plurality of forming roller sets which are arranged at intervals in the advancing direction. The plurality of forming roller groups are sequentially provided with a first side edge upright area, a first side edge inclined area and a first side edge parallel area along the advancing direction; the two ends of the upper forming assembly of the forming roller set located in the first side parallel area are each provided with a first side parallel wheel, the two ends of the lower forming assembly of the forming roller set located in the first side parallel area are each provided with a second side parallel wheel, and the first side parallel wheels and the second side parallel wheels correspond to each other to form a third forming gap. And when the profile plate passes through the third forming gap, the first inclined section of the profile plate gradually gets close to the main body of the profile plate, so that the first inclined section is parallel to the main body of the profile plate to form a U-shaped groove structure. According to the technical scheme, the forming equipment capable of forming the U-shaped groove plates on the two sides can be provided.

The invention belongs to the technical field of biomedical engineering and microfluidics, and particularly discloses a vascularized organ chip and method for large-scale culture, the vascularized organ chip comprises an upper runner layer and a lower culture layer, and the runner layer and the culture layer are fixedly sealed; a liquid storage pipe matched with the flow channel layer is adhered to the upper surface of the flow channel layer; according to the vascularized organ chip, the flow channel layer is used for pouring a culture solution to provide fluid stimulation and nutrients for cells; the three parallel-connected large-scale culture chambers of the culture layer are used for culturing cells, so that blood circulation of a microvascular system in a human body can be simulated, and a microenvironment in tissues can be reproduced more truly; meanwhile, the requirement for culturing a large number of tissue cells through the chip can be met, and the cell yield and the experiment scale can be effectively increased.

The invention discloses a photomask defect identification method, and belongs to the technical field of photomask detection. The method comprises the following steps: S1, acquiring a surface defect image; s2, preprocessing the surface defect image based on the imaging characteristics of the defect; s3, performing filtering algorithm correction on the preprocessed surface defect image; s4, performing colorimetric analysis on the surface defect image corrected by the filtering algorithm to obtain a particle region; s5, on the basis of morphological characteristics of defects, segmenting the particle region to obtain defect islands; and S6, performing morphological analysis on the defective island to obtain a defect type and a defect size. The photomask defect identification method provided by the invention combines morphological and colorimetric analysis methods to automatically process the surface defect image and extract the defect island, improves the accuracy of defect identification, gets rid of the constraint of the difference of different equipment imaging systems, realizes automatic batch processing, and saves the manpower and material investment in the mass production link.

The invention provides a flexible resource parameter identification method based on a KAN neural network, and relates to the technical field of power grid resource parameter identification, and the method comprises the following steps: obtaining historical operation data of power grid resources, and identifying flexible resources in the power grid resources through an analytic hierarchy process; according to the identified flexible resources, constructing a flexible resource digital model through a digital twinning technology, and extracting a sample data set according to the flexible resource digital model; based on the extracted sample data set, constructing and training a resource parameter identification model through a KAN neural network; and acquiring operation data of the flexible resources in real time, and identifying flexible resource parameters by using the trained resource parameter identification model. According to the method, the accuracy of identifying the flexible resource parameters by the model can be improved, more accurate decision support can be provided for power grid operators by identifying the flexible resource parameters, and more reasonable decisions can be made in the aspects of resource scheduling, energy distribution and the like.

The invention discloses a cross-platform operating system upgrade patch automatic deployment method and system, and relates to the technical field of operating system patch management, and the method comprises the following steps: analyzing a patch dependency relationship, and dynamically adjusting a deployment sequence; performing isolation deployment based on a containerization technology; and monitoring the state of each container in real time, and performing real-time monitoring and feedback optimization on a deployment result. By means of the method, accurate control over patch deployment in a complex heterogeneous environment is achieved, the failure rate caused by dependency conflicts is reduced, meanwhile, deployment time is shortened, stability and reliability of system operation are improved, independence between patches is ensured, and the service life of the system is prolonged. According to the method and the system, system crash caused by patch incompatibility or operation failure is avoided, the deployment efficiency is improved, the efficient container isolation and highly compatible patch deployment process is realized, the self-adaptability and long-term stability of the system are improved, and powerful technical support is provided for complex and changeable patch deployment scenes.

The invention relates to a virtual private network communication method, device and equipment and a readable storage medium, and the method comprises the steps: forwarding a to-be-transmitted data stream to a virtual private network gateway through a private network, and transmitting the data stream to a virtual private network channel through the virtual private network gateway, and transmitting the data stream to a client through at least two tunnels in the virtual private network channel. Compared with the prior art, the embodiment of the invention has the advantages that the private network and the virtual private network are separated, the routing independence of the private network is guaranteed, the to-be-transmitted data stream is further forwarded to the virtual private network gateway through the private network, and the data stream is sent to the virtual private network channel through the virtual private network gateway. And transmitting the data stream to the client through at least two tunnels in the virtual private network channel, using a plurality of VPN tunnels, providing active-standby and active-active modes, providing more VPN networking choices, and ensuring high availability of services.

The invention discloses an unmanned aerial vehicle hybrid deployment flight control system. The system comprises an airborne computing unit arranged on the unmanned aerial vehicle, a computing power CPU of the airborne computing unit comprises a plurality of cores, and flight control software is installed on the airborne computing unit in a hybrid deployment mode; a specific core of the computing power CPU runs the flight control software and the RTOS real-time domain; other cores of the computing power CPU run a Linux operating system and an intelligent computing domain, and the intelligent computing domain is used for executing an intelligent computing task; and different cores of the computing power CPU are connected through a communication bus. According to the invention, the flight control software and the intelligent computing task are mixed and deployed on different cores of the same airborne computing unit, so that when the airborne computing unit needs to be configured, an independent flight control board card is not used any more; flight control software and an intelligent computing task share one computing power CPU, redundancy and volume increase of multiple hardware modules are avoided, and installation complexity and cost are effectively reduced.

The invention provides an endoscope which comprises a main body and a photoelectric component, the main body comprises a first shell and a tube body, a mounting cavity is formed in the first shell, the tube body is connected to the front end of the first shell, and the interior of the tube body is communicated with the mounting cavity; the photoelectric assembly comprises a mounting box and an imaging unit connected to the mounting box, the mounting box is arranged in the mounting cavity, the imaging unit is arranged in the tube body, the mounting box comprises a housing, a connecting seat and an elastic part, a sealing cavity is formed in the housing, and a mounting hole is formed in the side wall of the housing; the connecting seat is arranged in the sealing cavity, and the outer wall of the connecting seat is in sealing connection with the mounting hole; the elastic piece is arranged in the sealing cavity, and the elastic piece is connected between the inner wall, facing the mounting hole, of the housing and the connecting base in an abutting mode. The invention is used for solving the technical problem that the photoelectric component in the endoscope is difficult to reuse.

The invention discloses a power grid consumption blocking factor identification method based on an RNN-SHAP model. The method comprises the following steps: determining a set of new energy consumption blocking factors according to measurement data and engineering experience; performing feature decorrelation by using a Pearson's correlation coefficient, establishing an RNN-SHAP model, taking data of a specific line model and section active power as input of a neural network model, taking the new energy blocking quantity as output of the neural network model, and predicting the new energy blocking quantity in different time periods through the RNN; the contribution degree of each blocking factor to the new energy blocking amount is analyzed by using an SHAP algorithm so as to identify the blocking key factor of each time period. According to the method, the RNN neural network and the SHAP algorithm are utilized, the influence degree of the blocking factors in each time period on the new energy blocking amount can be effectively estimated, and therefore the key factors causing new energy blocking in different time periods are determined.

The utility model provides a burner and a stove. The combustor comprises a fire distributor and an inner ring fire cover covering the fire distributor, a first gas mixing chamber and a second gas mixing chamber are defined between the inner ring fire cover and the fire distributor, and the first gas mixing chamber and the second gas mixing chamber are separated from each other. A first fire outlet hole communicated with the first gas mixing chamber and a second fire outlet hole communicated with the second gas mixing chamber are formed in the inner ring fire cover, and the second gas mixing chamber is eccentrically arranged relative to the inner ring fire cover. Thus, the problem that the first fire outlet hole and the second fire outlet hole compete with oxygen during combustion can be better avoided, sufficient combustion of gas at the first fire outlet hole and the second fire outlet hole is guaranteed, and the problems that smoke exceeds the standard, and heat efficiency is low are avoided.

The invention belongs to the technical field of power system dynamic assessment, and particularly relates to a machine learning-based urban power lifeline dynamic risk assessment method. The risk assessment method comprises the following steps: step S100, defining a power system risk assessment problem; step S200, the data input layer performs multi-source heterogeneous data fusion, and collects and integrates multi-source heterogeneous data; s300, a feature extraction layer extracts risk features capable of reflecting risks of the power system from the integrated multi-source heterogeneous data, and optimizes the risk features through an automatic feature selection mechanism to form feature vectors; step S400, the model training and evaluation layer uses the optimized risk features to train and debug a risk evaluation model, and evaluates the performance of the model; and step S500, dynamically displaying a risk assessment result by a result output layer. The risk assessment method has enough training samples, can overcome the defects of training data scarcity, difficult feature optimization and insufficient model generalization ability, and realizes accurate identification and dynamic assessment of urban power system risks.

The invention belongs to the technical field of power system dispatching, and discloses a constraint learning embedded optimization method and system for an AC security constraint unit committee, and the method comprises the steps: constructing a mixed integer nonlinear model of an AC power flow-considered SCUC problem of a to-be-optimized power system; constructing a data set by taking the input data of the alternating current power flow constraint as a sample and taking the output data of the corresponding alternating current power flow constraint as a label; adopting the data set to train a neural network in an off-line manner to obtain a trained neural network representing the alternating current power flow constraint; linearizing a neural network representing the alternating current power flow constraint, and embedding the linearized neural network into the mixed integer nonlinear model to replace the alternating current power flow constraint in the mixed integer nonlinear model; and solving the embedded model on line by adopting an optimization algorithm to obtain an optimal solution of the security constraint unit commitment problem of the to-be-optimized power system considering the alternating current power flow. According to the method, the solving precision and solving efficiency of the AC power flow SCUC problem can be improved.

The invention relates to the technical field of carbon emission prediction and early warning, and provides a multi-scale neural distribution prediction and hierarchical migration early warning method, which comprises the following steps: acquiring historical carbon emission data, respectively inputting a historical sequence and a to-be-predicted sequence into an energy consumption stochastic differential equation model and a carbon factor stochastic differential equation model, generating a multi-scale carbon emission path sample set through an independent random disturbance term; calculating a path-level suitability score of each path sample based on the standard-exceeding risk integral, the first standard-exceeding moment and the path fluctuation variance; layering the calibration data set into a plurality of working condition layers according to working condition labels, sharing distribution shape parameters among the working condition layers through a hierarchical Bayesian method, and regularizing quantiles of small sample working condition layers to obtain an early warning threshold value of each working condition layer; and selecting a corresponding early warning threshold value according to the current working condition label to compare and trigger early warning. According to the method, the accuracy of carbon emission distribution prediction and the robustness of an early warning system are improved, and the problem that the early warning threshold value is unstable under the small sample working condition is relieved.

The utility model discloses an adjustable flushing aspirator, which belongs to the technical field of medical instruments and comprises a handle and a tube component, an anterior chamber with an opening at the front end, a first channel and a second channel are arranged in the handle, and the first channel and the second channel are communicated with an inner cavity of the anterior chamber and isolated from each other. The tube assembly comprises a suction tube arranged in an inner cavity of the anterior chamber and connected with the first channel and a flushing tube connected to the outer side of the suction tube in a sleeving mode and connected with the anterior chamber, the second channel is communicated with a radial gap between the suction tube and the flushing tube through an annular gap formed between the suction tube and the inner cavity of the anterior chamber, and the second channel is externally connected with a flushing connecting tube. The flushing connecting pipe is provided with a flow adjusting assembly used for adjusting the communication section area of the flushing connecting pipe and the second channel, and the surface of the handle is provided with a pressure adjusting hole communicated with the first channel. According to the utility model, the suction channel and the flushing channel can be separated from each other, and the flushing flow and the negative pressure can be adjusted as required.

The invention relates to a plug seedling pretreatment device. The device comprises a transverse conveying mechanism, a longitudinal conveying mechanism, a reversing mechanism and a trimming mechanism, the transverse conveying mechanism and the longitudinal conveying mechanism are perpendicular to each other, and the transverse conveying mechanism and the longitudinal conveying mechanism are connected through the reversing mechanism; the transverse cutting mechanism is mounted on the transverse conveying mechanism and is used for transversely cutting the seedlings in the hole tray; the longitudinal cutting mechanism is mounted on the longitudinal conveying mechanism and is used for longitudinally cutting the seedlings in the hole tray; the trimming mechanism is mounted on the longitudinal conveying mechanism and is used for trimming the cut seedlings, so that the heights of the seedlings are kept consistent. According to the scheme provided by the invention, the seedlings in the hole tray can be pretreated, the phenomena of entanglement and adhesion among the seedlings are prevented, and subsequent transplanting is facilitated.

The present invention provides a high-speed satellite telemetry processing method and system, comprising the following steps: segmenting the original telemetry data file according to the number of computer CPU cores, and evenly dividing it into N primary data segments; sliding and searching for a telemetry frame independent of the previous frame at the starting position of the primary data segment, and re-segmenting the file with the starting position of the independent telemetry frame as the segmentation point to obtain N secondary data segments; the main thread newly opens N processing sub-threads, and allocates the i-th secondary data segment to the i-th thread; each processing sub-thread synchronously performs telemetry processing and parsing tasks; the main thread waits for all processing sub-threads to complete their respective telemetry processing tasks, and then splices the telemetry processing result data in sequence to generate the final telemetry processing result data file. The present invention performs segmentation and splicing processing on telemetry data, solves the problem of segmentation of cross-frame packets in the packetized telemetry scheme, and ensures the integrity and balance of the data, and the independence of the segmented data.

This disclosure provides a method for early warning of safety and environmental protection for large-scale industrial solid waste stockpiles based on digital twins. The method includes: acquiring the stockpile state variables within the target stockpile and constructing a comprehensive stockpile state set with a unified time reference based on these variables. Based on the comprehensive stockpile state set, a dual-state mapping structure is constructed in the digital twin space to obtain a dual-state digital twin stockpile model. The dual-state digital twin stockpile model is used to analyze the target stockpile area to identify its mechanical state characteristics and determine the critical instability pre-state based on these characteristics. The critical instability pre-state is used to determine whether the target stockpile area is in a critical state of mechanical degradation. When the target stockpile area is in a critical state of mechanical degradation, safety and environmental protection early warning information is output, enabling early warning of instability and pollution risks in high-moisture-content fine-particle solid waste stockpiles, thus improving the reliability and timeliness of safety and environmental protection risk assessment.

The invention discloses a control method for an electronic hydraulic braking system, and relates to the technical field of vehicle braking, and the control method comprises the step of controlling opening and closing of a decoupling electromagnetic valve, a first linkage electromagnetic valve and a second linkage electromagnetic valve based on a braking mode instruction so as to execute a braking action corresponding to a wheel braking unit. By the adoption of the control method for the electronic hydraulic brake system, the brake master cylinder is arranged to be of a three-cavity structure, brake control over the wheel brake unit under multiple hydraulic lines is achieved through pressure adjustment of the three brake cavities of the brake master cylinder based on opening and closing control over the decoupling solenoid valve, the first linkage solenoid valve and the second linkage solenoid valve, and the brake control over the wheel brake unit under multiple hydraulic lines is achieved. The safety redundancy capacity of the braking system is greatly improved, and the braking stability under different working conditions is guaranteed.

The embodiment of the invention relates to the technical field of oil and gas reservoir engineering, and discloses an online filling and reporting method, system and device for oil and gas data and a storage medium. The method comprises the steps that corresponding data filling and reporting tables are generated according to different oil and gas data; a corresponding storage process is set for each data filling table, so that each kind of oil and gas data is automatically and structurally stored in a database according to the corresponding storage process; managing the filling permissions according to the user attributes, and associating the filling permissions corresponding to the user attributes with the data filling table; and when data filling is carried out, according to the filling permission corresponding to the attribute of the currently logged-in user, matching a corresponding data filling table for the currently logged-in user to fill. By utilizing the exemplary embodiment of the invention, the difference between the stored data and the original data can be quickly compared, the error checking and correcting cost is greatly reduced, and finally, the safety management, efficient storage and high-quality filling of the data are realized.

Provided is an image encoding / decoding method and device. The image decoding method disclosed herein may include the following steps: obtaining maximum transform block size information; deriving the maximum transform block size based on the maximum transform block size information; and determining whether to apply a secondary inverse transform to the current block based on the maximum transform block size and the horizontal size and vertical size of the current block.

This application relates to a high-precision modeling method for low-Earth orbit (LEO) satellite orbits based on measured data correction. The method includes: first, acquiring measured orbital error data of in-orbit LEO satellites; setting the simulation scenario duration; periodically extending or pruning / segmenting the error data; generating a pseudo-random sequence using a 10-level feedback shift register; modulating and combining the pseudo-random sequence to obtain a first and a second sequence; filtering the valid second sequence through correlation calculation to obtain uncorrelated sequences for each segment; finally, combining and sampling according to the number of satellites, adding the orbital error to the error-free orbit to achieve high-precision modeling. This method can accurately reflect the characteristics of measured orbital errors, achieving high-precision modeling of LEO satellite orbits.

The invention relates to the technical field of vehicle navigation, and discloses an airborne closed cabin vehicle path dynamic navigation planning method, which comprises the following steps: S1, navigation initialization, S2, real-time perception data acquisition and correction, and upward navigation reference based on a target established in the step S1, S3, map-free error correction and fusion, S4, dynamic path planning, and S5, accurate correction before landing. And S6, outputting a navigation control instruction based on the real-time navigation track in the step S4. Complete autonomous navigation of an airborne closed cabin vehicle in the whole process is achieved, dependence on external visual reference, a pre-built map or a ground guide signal is not needed, an absolute direction anchor point is provided through a target upward navigation benchmark, multi-source sensor decoupling correction and adaptive Kalman filtering fusion are combined, airflow disturbance compensation is synchronously brought in, and the navigation precision is improved. The inertial navigation accumulative error is effectively eliminated, the method can be adapted to an extreme airborne environment without reference, map and signal, and the independence and reliability of the navigation process are guaranteed.

The invention discloses a hydraulic oil return link valve body which comprises a cylindrical shell internally provided with a cavity, a plurality of oil inlet pipes penetrating into the shell are evenly distributed on the periphery of the shell, and each oil inlet pipe is tangent to the side wall of the shell; an oil return device is arranged in each oil inlet pipe and used for guiding hydraulic oil into the shell; an oil outlet pipe is arranged at the top of the shell and is used for discharging the treated hydraulic oil; a rotating disc is arranged in the shell and located below all the oil inlet pipes in a rotating mode, a plurality of blades are evenly distributed on the rotating disc, an independent space is formed between every two adjacent blades and used for separating and guiding hydraulic oil from different oil inlet pipes, and it is guaranteed that the oil return processes of the oil inlet pipes do not interfere with one another. Compared with the prior art, the oil return process of the hydraulic actuator is effectively managed, it is ensured that oil return of the oil inlet pipes does not interfere with one another, and meanwhile oil is purified and discharged smoothly.

The application provides a unified evaluation method and device for a multi-control paradigm video world model, wherein the method comprises the following steps: based on unified case metadata of a to-be-evaluated model, an evaluation request carrying a control condition is sent to a unified service interface, and the unified service interface distributes the evaluation request based on a deployment mode of the to-be-evaluated model; and an evaluation video generated by calling the to-be-evaluated model is received by a remote server or a local server. By using unified case metadata and a unified service interface, the evaluation request is distributed to the remote server or the local server according to the deployment mode of the to-be-evaluated model, a unified evaluation entry is provided for various video world models of a control interface, the local server is designed as an isolated resident inference service started by a sub-process, the independence of the evaluation environment of each open source model is ensured, the model weight is loaded once when the service is started, repeated loading of a huge model weight is avoided, and the expansibility and automation level of the evaluation system are improved.