2841 results about "Equipment computers" patented technology

The present application provides training and speech generation methods and apparatuses for a speech generation model, an electronic device, a computer-readable storage medium, and a computer program product. The method comprises: obtaining a first speech generation model; obtaining sample data of a plurality of modalities; on the basis of a prompt image sequence and speech text, respectively calling a plurality of encoders to perform encoding, so as to obtain a multi-modal encoding vector sequence; on the basis of the multi-modal encoding vector sequence, calling a decoder to perform decoding, so as to obtain decoded text; determining a probability distribution for the decoded text and the sample data of the plurality of modalities, and determining a target loss on the basis of the probability distribution; and on the basis of the target loss, updating parameters of the decoder and at least one of the encoders, wherein the updated decoder and the plurality of updated encoders are configured to form a second speech generation model, and the second speech generation model is used to generate target speech text corresponding to a prompt image sequence of a target object.

The invention relates to a multi-agent cooperative recruitment method, system and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the steps of obtaining a target recruitment demand, and generating a task tree for the target recruitment demand through a preset dynamic decision-making mechanism in a main agent; determining a time sequence relationship and logic constraint information of each sub-task in the task tree, and calling target sub-agents in the sub-agents in sequence based on the time sequence relationship and the logic constraint information; the target sub-agent is used for executing the sub-task in the task tree; and generating a target recruitment task result until task execution results of the sub-tasks are obtained. By adopting the method, the recruitment efficiency can be improved.

Provided in the embodiments of the present disclosure are an optimal control method and system for a distribution network with a high penetration of distributed generators, and an electronic device, a computer-readable storage medium and a computer program product. The optimal control method is applied to a centralized control platform. The method comprises: on the basis of an association relationship between power outputs of distributed generators at various levels, performing multi-level forecasting on historical power generation data of each distributed generator, so as to obtain a day-ahead forecasted output value of each distributed generator; for each distributed generator, determining a day-ahead output adjustment value on the basis of historical power flow data and an optimal power flow optimization model; and sending to the corresponding distributed generator the output adjustment value and forecasted output value of each distributed generator as centralized control instructions, so that the distributed generator controls an output value during each control period within the day ahead on the basis of an acquired real-time output measurement value, the acquired output adjustment value and the acquired forecasted output value.

The invention relates to a memory management method and device applied to a server-free data center, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: in response to a virtual memory application instruction, performing virtual memory allocation in a local computing resource pool to obtain a virtual address space; in response to the virtual address access instruction, acquiring cache information of the computing resource pool, and triggering missing page interruption under the condition that data corresponding to the virtual address is not hit in the cache information; the cache information comprises locally cached cache data and page table information; acquiring physical memory resources in a memory resource pool according to the virtual address and the page table information under the condition of triggering missing page interruption of the anonymous page; and updating the page table information and the cache information according to the physical memory resources. By adopting the method, the processing delay of the data center can be reduced while the decoupled memory management logic is used for improving the resource utilization rate.

The invention relates to a security authentication method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: receiving a key generation request sent by target equipment; if it is determined that the key generation request is legal, generating a trust identifier and a security certificate of the target device, obtaining an identifier verification request through the trust identifier, and sending the identifier verification request to an authentication server, so that the authentication server verifies the trust identifier to obtain an access token corresponding to the target device; receiving a message returned by the authentication server, and processing an access token carried by the message to obtain a trust identifier; and writing the security certificate into the target protection area, and sending the trust identifier and the security certificate to the target equipment, so that the target equipment is registered in the authentication server. By adopting the method, the identity verification of the equipment and the control of an encryption communication mechanism are realized, and the safety protection performance in a communication system is further improved.

The invention discloses a micro burst traffic detection method, network element equipment, a computer readable storage medium and a computer program product, and relates to the technical field of communication, the method is applied to the network element equipment, the network element equipment is provided with a microsecond-level timer, and the method comprises the following steps: counting the number of forwarding message bytes of each sub-time window in a sliding time window; performing mean value calculation on the forwarding message byte numbers of all the sub-time windows in the sliding time window to obtain a sub-window byte number mean value of the sliding time window; determining a sub-window byte number threshold value of the sliding time window based on the sub-window byte number mean value; and determining the state of the sub-time window of which the forwarding message byte number is greater than the byte number threshold value of the sub-window as a flow micro-burst state. According to the invention, the size of the micro-burst flow and the corresponding time delay can be accurately detected in a fine-grained manner, and the detection precision of the micro-burst flow is effectively improved.

The invention relates to a KV cache optimization method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: calculating a key vector and a value vector corresponding to each element in a text input sequence input into a large language model; through a multi-head potential attention mechanism, performing low-rank joint compression on the key vector and the value vector to obtain a potential vector, and storing the potential vector in a KV cache space; based on a scaling law, determining an optimal compression dimension, regenerating an adaptive potential vector and updating a KV cache space; for the same text input sequence, generating corresponding query vectors, and grouping the query vectors according to a preset grouping rule; calculating a semantic association weight between each group and the correspondingly called potential vector, and taking the semantic association weight as a group attention calculation result; in the reasoning process, potential vectors and grouping attention calculation results are calculated to calculate attention weights. By adopting the method, the storage requirement of the KV cache can be further reduced.

The invention relates to a large model energy consumption optimization method and device, computer equipment, a computer readable storage medium and a computer program product. Comprising the steps of collecting hardware state data and task data of a large model; performing stage identification according to the task data, and determining a current stage; performing state prediction according to the hardware state data and the task data through a prediction model corresponding to the current stage to obtain target state information; generating optimization parameters of the current stage through a joint optimizer according to the target state information; and adjusting the resources of the large model according to the optimization parameters. In combination with stage perception and dynamic resource adjustment, a differentiated resource adjustment strategy based on different stages is realized, specific energy consumption pain points in an AI scene are solved, high power consumption of large model training, instantaneous fluctuation of reasoning requests and the like are reduced, sustainable and efficient operation of an AI system is ensured, the computing power demand and resource consumption of a large model are balanced, and the system performance is improved. And wide landing and development of a large model in various scenes are facilitated.

The invention relates to a data processing method and device, computer equipment, a computer readable storage medium and a computer program product. After first matrix data is obtained, tensor data to be processed are obtained, and the tensor data are distributed to a plurality of calculation units; calculating a local feature value corresponding to the target calculation unit according to the local tensor data allocated to the target calculation unit; dividing the target calculation unit into data blocks corresponding to the number of the plurality of calculation units; according to the serial number of the data block, mapping the local feature value recorded in each data block to a target storage block of a target memory until the local feature value in each calculation unit is distributed to the corresponding storage block in the target memory, and obtaining a global feature value table; according to the global feature value table, feature value calculation is carried out to determine the target feature value, conflict-free parallel calculation is achieved, the utilization rate of calculation resources is increased, and the data processing efficiency is improved.

The invention discloses a cerebral hemorrhage image recognition method, and aims to solve the problems that different cerebral hemorrhage focuses have significant differences in form, size and position, and cerebral anatomical structures of different individuals also have significant differences, so that the features of cerebral hemorrhage images are complex and changeable, and the recognition accuracy is high. Therefore, a traditional image processing method is difficult to accurately identify the key information of the focus in the cerebral hemorrhage image. The method comprises the steps of obtaining a cerebral hemorrhage image; the cerebral hemorrhage image is input into the trained cerebral hemorrhage image recognition model to obtain a recognition result output by the model, and the recognition result comprises the lesion type, position and size of the lesion in the cerebral hemorrhage image; wherein the cerebral hemorrhage image recognition model is obtained by training an enhanced cerebral hemorrhage image sample obtained by performing elastic deformation processing on a cerebral hemorrhage image sample. The invention further discloses a cerebral hemorrhage image recognition device, electronic equipment, a computer readable storage medium and a computer program product.

The invention provides a text reinforcement learning method and device, electronic equipment and a computer storage medium, relates to the technical field of text reinforcement learning, and is applied to a text generation model. The method comprises the following steps: determining one or more target vocabularies in a statement to be analyzed, and generating a replacement word set according to the target vocabularies; replacing the target vocabulary in the to-be-analyzed statement with a current candidate replacement word to obtain a replacement statement; calculating one or more difference index values between the replacement statement and the to-be-analyzed statement; wherein the difference index value is used for quantifying the influence of the replacement of the target vocabulary on the original sentence from different dimensions; based on the difference index value, determining an importance weight of the target vocabulary in the to-be-analyzed statement; wherein the importance weight is used for redistributing the reward of the single character in the sentence. Vocabulary importance is quantified through disturbance analysis and multi-dimensional evaluation, and then fine-grained distribution of rewards is achieved.

The invention provides a virtual resource processing method and device, electronic equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: in response to a customer, selecting a payment mode for a merchant as a target payment mode, jumping from a merchant platform to an acquirer platform corresponding to the target payment mode, and displaying a payment order in the acquirer platform; in response to a selection operation for a payment bank, displaying collection account information of the merchant; in response to a copy operation for the payment account information, skipping to a payment page of a payment bank from the acquirer platform, and displaying collection account information in the payment page; and in response to payment operation triggered based on the payment page, transferring the virtual resource of the payment amount from the payment account to a transition account associated with the collection account, and transferring the virtual resource of the payment amount from the transition account to the collection account under the condition that the risk identification of the acquirer platform on the payment account is passed. According to the invention, the security of the collection account can be improved.

The invention discloses a control method for steam condensate waste heat recovery and waste water heating, a control device for steam condensate waste heat recovery and waste water heating, control equipment and a computer readable storage medium. The thermal load and the logarithmic average temperature difference are calculated based on the energy conservation principle, and the needed heat exchange area is calculated through a logarithmic average temperature difference method. A flow combination scheme of the plate heat exchanger is determined according to the heat exchange area and the flow parameters, a steam condensate side adopts a single flow, and a waste water side adopts asymmetric configuration of double flows or multiple flows so as to adapt to the flow difference of the two sides. In addition, a real-time monitoring mechanism is further established, when the total heat transfer coefficient is decreased by more than 10%, a cleaning program is automatically triggered, continuous and stable operation of the system is guaranteed, and efficient utilization of waste heat resources is achieved.

The invention relates to a cache data processing method and device, computer equipment, a computer readable storage medium and a computer program product. The method is applied to a distributed storage system adopting a three-layer architecture design, and the three-layer architecture comprises a service application layer, a data processing layer and a persistence layer. Comprising the following steps: acquiring access request monitoring information through a service application layer to obtain an access feature data set, and performing pattern classification identification on the access feature data set to obtain a predicted access pattern; generating a prefetching task list and prefetching data according to the predicted access mode, and storing the prefetching data to a data processing layer for caching processing; and executing each prefetching task according to the sequence of the prefetching task list, and reading the prefetching data from the mechanical hard disk and writing the prefetching data into a cache space in a solid state disk resource pool under the condition that the prefetching data does not exist in the solid state disk resource pool of the persistence layer. By adopting the method, the cache performance can be improved.

The invention relates to a robot autonomous navigation method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: acquiring a plurality of sensing data sets from different sensors; wherein the sensing data sets carry the same timestamp; generating a semantic map and a geometric map of the target area according to the sensing data sets; acquiring voice control information, and matching at least one semantic tag in the semantic map according to the voice control information; and generating a navigation path according to the matched semantic tag and the geometric map, and completing an autonomous navigation task of the robot according to the navigation path. The method can improve the autonomous navigation stability of the robot.

The invention provides a battery anti-theft method, which is applied to a battery management system (BMS) and comprises the following steps: receiving a first state detection command sent by a centralized monitoring unit (CSU); obtaining battery pack change information of a battery pack where a target battery to which the BMS belongs is located from the BMS; determining the group state of the battery pack according to the battery pack change information, and locking the target battery in response to the change of the group state of the battery pack; according to the embodiment of the invention, the state change of the battery group is detected by using the intelligent bionic algorithm, the target battery causing the state change of the battery group is automatically identified, the target battery is locked to realize theft prevention, and the battery theft prevention cost and the detection accuracy can be considered at the same time. The invention further provides electronic equipment, a computer readable medium and a computer program product.

The invention relates to a power distribution network dispatching optimization method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: acquiring power supply output information and load demand information of a plurality of polymers in a power distribution network; according to the power supply output information and the load demand information, determining reference interaction power among the plurality of polymers; the reference interaction power is optimized according to the parameter information of the power supply devices in the polymers through the polymers until a preset convergence condition is met, and target interaction power among the multiple polymers is obtained; according to the target interaction power and the parameter information of the power supply equipment in each aggregate, target output information of the power supply equipment in each aggregate is determined; and performing energy scheduling on the power supply equipment in each polymer according to the target output information. By adopting the method, the dispatching optimization efficiency and effect of the power distribution network can be improved.

The invention relates to a message processing method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: in response to an RDMA request of a sending end, obtaining an RoCE message corresponding to the RDMA request; converting the QP pair information in the RoCE message into a corresponding hash operation result; and generating an RDMA flow identifier of the RoCE message according to the Hash operation result. By adopting the method, the load balancing effect can be improved.

Provided are a robot control method and apparatus, an electronic device, a storage medium, and a program product. The method includes: acquiring state data configured for indicating a current motion state of a robot, and acquiring environmental data configured for indicating an environment where the robot is currently located; predicting, based on the state data, an initial action parameter configured for controlling the robot to imitate an object action of a target object; predicting, based on the environmental data, an environmental impact parameter configured for representing impact generated by the environment on imitation of the object action by the robot; fusing the initial action parameter and the environmental impact parameter to obtain a fused action parameter of the robot; and generating a control instruction based on the fused action parameter, the control instruction being configured for controlling the robot to perform a target action indicated by the fused action parameter.

The invention provides a vehicle drifting control method, device and equipment, a computer readable storage medium and a computer program product. The method comprises the steps that based on target state information and actual state information of a vehicle, a first slip rate component and a second slip rate component are obtained, the first slip rate component is used for slip angle stability control, and the second slip rate component is used for yaw velocity control; determining a left rear wheel slip rate target and a right rear wheel slip rate target based on a basic slip rate target, the first slip rate component and the second slip rate component; and rotating speed calculation is conducted based on the actual state information, the left rear wheel slip rate target and the right rear wheel slip rate target, and the vehicle is controlled based on the obtained left rear wheel rotating speed target and the obtained right rear wheel rotating speed target. According to the invention, a driver can maintain stable drifting of the vehicle only by operating the steering wheel and the accelerator.

The invention relates to a policy control security enhancement method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: acquiring a policy request initiated by an external entity to a policy control plane in a cross-multi-cloud or cross-domain environment; extracting multi-dimensional features from request data carried in the strategy request, and constructing a multi-dimensional feature map according to the multi-dimensional features; performing anomaly detection on the strategy request according to the multi-dimensional feature map to generate a threat judgment result; under the condition that the threat judgment result shows that the strategy request is normal, performing semantic, resource and authority triple verification on the strategy request to generate a verification judgment result; and under the condition that the verification judgment result shows that the strategy request is verified successfully, issuing the strategy to the target execution point and enabling the strategy to take effect, thereby completing strategy life cycle management. By adopting the method, the security robustness, privacy compliance and cross-domain cooperation efficiency of the strategy control system can be improved.

The present application relates to the technical field of speech processing, and particularly relates to a mixed speech processing method, an electronic device and a computer readable medium. The method comprises the following steps: collecting mixed speech and environmental influence parameters; performing bionics frequency domain analysis on the mixed speech to obtain low-frequency attenuation compensation feature data; performing multipath effect propagation analysis on the low-frequency attenuation compensation feature data through the environmental influence parameters to generate channel distortion data; performing time domain-frequency domain joint deconvolution processing on the mixed speech by using the channel distortion data to generate direct sound components and reflected sound components; performing adversarial training based on the direct sound components and the reflected sound components to generate anti-multipath speech enhancement data; and constructing a dynamic frequency compensation filter based on preset environmental acoustic characteristics. The present application improves the output quality of mixed speech through multi-stage signal processing, frequency compensation and real-time optimization technology.

A device, computer readable medium and method for waking up a network device, involving user equipment, a first network device, and a second network device. The method includes determining and transmitting a wake-up signal from the user equipment based on transmit power, which is dynamically adjusted using communication protocol parameters such as initial transmit power, step increments, and maximum power thresholds. Wake-up signals are transmitted multiple times within a set period to ensure reliable reception. The first network device processes the wake-up signal, assesses signal quality, and interacts with the second network device to determine wake-up conditions based on agreed communication protocols and signal quality thresholds. The second network device manages wake-up coordination among multiple first network devices, identifying devices meeting quality conditions and transmitting wake-up indications and configuration details.

The invention relates to a fault data processing method and device based on a large model, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: acquiring target fault record data in a target time period; removing irrelevant contents in the target fault record data through a preset large language model and target knowledge base data to obtain key fault information, and performing attribution analysis on the target fault record data to obtain a fault reason; and classifying the fault phenomena and the fault causes to obtain a fault phenomenon classification standard and a fault cause classification standard. By adopting the method, the original fault report data can be automatically analyzed and sorted in combination with a large language model and a knowledge base plan, so that automatic data classification and information extraction are realized, manual operation is avoided, and efficient classification and analysis of fault data are realized.

The embodiment of the invention provides a content display method and device, equipment, a computer readable storage medium and a product. The method comprises the steps that a first display interface associated with theme content is displayed; displaying a first presentation result in the first display interface, the first presentation result being generated by a first model based on a plurality of second media contents associated with the subject content, and the first presentation result at least comprising a plurality of text paragraphs; and presenting a first media content or an association identifier at the association position of the plurality of text paragraphs, wherein the first media content or the association identifier is associated with the first model. Therefore, the availability of the first presentation result can be increased, and the interaction mode for the first presentation content can be enriched. In addition, the content type of the display content in the first display result can be enriched, and the content quality of the first display result is improved.

The invention provides an audio processing method and device, electronic equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: acquiring a structured list pre-generated for an audio to be edited; analyzing the structured list to obtain metadata of each audio clip; in response to an editing operation for the first audio clip, updating the metadata of the first audio clip to obtain updated metadata; obtaining the first audio clip based on the index of the first audio clip, and decoding the first audio clip to obtain corresponding audio data; creating an audio buffer source node, configuring playing parameters of the audio buffer source node according to the updated metadata, and calling the configured audio buffer source node to play audio data; and in response to receiving the file export request, constructing an offline audio context based on the updated metadata, and generating an audio file through the offline audio context. According to the invention, the audio editing efficiency can be improved.

Embodiments of the present application provide an image processing method and apparatus, an electronic device, a computer-readable storage medium, and a computer program product, which are at least applied to the field of artificial intelligence field or the field of image generation. The method comprises: acquiring a description text, and a raw image having a target object; encoding the raw image, to obtain an image encoding feature; encoding the description text, to obtain a text encoding feature; performing multi-level image residual mapping on the image encoding feature, to obtain an image semantic identification feature of the target object under each level among multiple levels; on the basis of the image semantic identification feature under each level, performing multi-level text residual mapping on the text encoding feature, to obtain a fused feature, the number of levels of the multi-level text residual mapping being the same as the number of stages of the multi-level image residual mapping; and decoding the fused feature, to obtain a generated image, the generated image comprising an object feature of the target object and a description feature of the description text.

The invention discloses an anti-offset wireless charging method, which comprises the following steps of: constructing a transmitting coil array as a charging end, detecting a relative position between a receiving coil of equipment needing wireless charging and the charging end, calculating offset data of the receiving coil, and taking a current state of each sub-coil in the transmitting coil during charging as a decision variable; constructing a dual-objective optimization mathematical model taking the minimum mutual inductance deviation and the maximum system transmission efficiency as optimization objectives, solving the dual-objective optimization mathematical model by adopting a genetic algorithm, searching to obtain a Pareto optimal solution set, selecting the Pareto optimal solution set, and controlling the current state of a transmitting coil in a charging end; through a decision variable control mechanism based on the sub-coil current state, the control granularity is refined from the whole coil unit to the sub-coil level, so that the system can dynamically adjust the current distribution mode in the unit according to the tiny offset of the receiving coil, local fine tuning of the magnetic field intensity is realized, and the precision of the magnetic field intensity is improved. And the coupling requirement under the complex offset working condition can be effectively met.

The invention discloses a lane-level congestion identification method, apparatus and device, and a computer program product. The method comprises the steps of receiving roadside sensing data and determining a sensing result of a vehicle target according to the roadside sensing data; dividing the perception result of the vehicle target to the corresponding lane to obtain the perception result of the vehicle target of each lane; based on a pre-created sliding window, respectively carrying out statistical analysis on the perception results of the vehicle targets of each lane to obtain the average speed of each lane; and determining lane-level congestion information according to the average speed of each lane and the road speed limit information. The vehicle sensing result is obtained by means of the strong sensing capability of the roadside equipment, the sliding window is used for carrying out statistical analysis on the data of each lane, and the lane congestion condition is accurately judged in combination with the road speed limit information for reference use of the vehicle, so that the problems of limited view and insufficient sensing capability of the vehicle are solved, and the driving safety is improved. Therefore, timely and accurate lane-level congestion information is provided for the automatic driving vehicle, and the driving efficiency and safety are improved.

Disclosed in the present application are a vehicle-mounted panoramic image generation method, a computer device, a computer storage medium, a computer program product and a mobile platform. The vehicle-mounted panoramic image generation method comprises: acquiring in real time multiple pieces of image data collected by a plurality of cameras; determining point cloud data corresponding to the multiple pieces of image data; on the basis of the point cloud data, correcting a preset three-dimensional model to obtain a corrected three-dimensional model; on the basis of the multiple pieces of image data, performing texture mapping on the corrected three-dimensional model to obtain a dynamic real-scene model; and on the basis of vehicle information and the dynamic real-scene model, generating a vehicle-mounted panoramic image. In the present application, point cloud data corresponding to multiple pieces of image data acquired in real time is used, a preset three-dimensional model is corrected on the basis of the point cloud data, and the corrected three-dimensional model is used to perform texture mapping to obtain a dynamic real-scene model, thereby solving the problem in the prior art whereby ambient objects with height are deformed and stretched due to a mismatch between a preset three-dimensional model and object shapes.