7 results about "Loss network" patented technology

The application belongs to the technical field of computer vision, and provides a rapid image style conversion method based on deep learning, comprising: constructing a double-network architecture through a generation network and a loss network; the loss network receives an original image, a reference image and a generated image, extracts features and calculates a perceptual loss to provide gradient information; when training a specific style model, input a single original image into the generation network to obtain a generated image; then input the three types of images into the loss network, extract content and style loss features and calculate a perceptual loss; adjust the generation network weight by minimizing the loss through gradient descent; iterate the data set for multiple rounds until the perceptual loss converges and the parameters are stable, and then the specific style model is obtained; when used, the original image is cropped, rotated and de-meaned to be standardized, and then input into the model generation network to output a specific style image; the application can realize the conversion through the double architecture of "generation network + loss network", and only one model training is needed to complete the conversion through one-time forward propagation of a single image.

Methods are disclosed for UE handling of arrangements of AI / ML related measurements and data collected but not yet transferred to the network in case of termination or loss of network connection. This can include mobility operations (e.g., handover); RRC state transitions; radio link failure; or change of PSCell. From configuration information received from a source network node specifying measurement and data collection for AI / ML model training, the UE handles the AI / ML data according to one of several procedures. The UE can delete it; send it to a target RAN node (with which the UE reestablishes network connection); continue performing AI / ML measurements; indicate to the target RAN node availability of previously collected data; or indicate to the target RAN node whether the source node configured AI / ML measurement configuration.

The invention provides a scene digital evaluation method and electronic equipment, and relates to the technical field of deep learning, and the method comprises the steps: interactively obtaining the design data of a transformer substation, and executing the point cloud time series data collection of the transformer substation based on a laser radar; extracting a stable static structure entity according to the design data, and establishing an initial anchor point candidate region; after the coordinates of the time sequence point cloud data set are reconstructed according to the collection coordinates and the collection parameters of the laser radar, loss identification under a time sequence consistency loss network is carried out; component semantic recognition is carried out on the time sequence point cloud data set, and a standardized component template is called; and after substation scene partition fusion, performing supplementary fusion by using a standardized component template, and performing evaluation management by using substation digital modeling. According to the method and the device, the technical problem of low substation scene digital modeling precision in the prior art can be solved, the technical target of improving the substation scene digital modeling precision is achieved, and the technical effect of improving the equipment identification accuracy is achieved.

The invention relates to a zero-packet-loss network implementation and verification method based on a Loongson CPU and an operating system, and belongs to the technical field of communication. In order to solve the problems of poor reliability, easy packet loss and high cost of a PRP switch of a traditional network, a Ubuntu system is installed, and a Loongnnix operating system cross compiling environment is installed; the method comprises the following steps: creating a PRP virtual network device under a Loongnnix operating system, and adapting a PRP protocol stack code to the PRP virtual network device; building a hardware environment of the PRP network based on the Loongson board card and a common switch; programming the compiled operating system on a board card, and connecting and configuring a PRP function in the board card through a PC; the PRP network is verified through a board card ping PC (Personal Computer), board card mutual ping and a network cable disconnection test mode. According to the invention, the PRP protocol stack is adapted to the Loongson board card, so that a zero-packet-loss network can be realized by using a common switch.

The invention relates to the technical field of power transmission network online edge communication, in particular to a power transmission network online edge communication method. According to the method, meteorological data and a network topology space are superposed, threatened nodes are identified, alternative paths are pre-calculated, service migration and gateway role transfer are actively completed before weather comes, and a robust communication path is established. When the network is divided into islands by extreme weather, the key monitoring data is ensured not to be lost by electing temporary sink nodes and implementing local hierarchical storage. And after the network is recovered, orderly return of historical backlog data and real-time data is realized by adopting a metadata abstract exchange and bandwidth negotiation mechanism, and channel congestion is avoided. According to the method, self-adaptive survival and flexible operation of the network under severe conditions are realized, and complete delivery of monitoring data, especially key alarms, is ensured.

The application provides a scene digital evaluation method and electronic equipment, and relates to the technical field of deep learning. The method comprises the following steps: obtaining design data of a transformer substation through interaction, and performing laser radar-based transformer substation point cloud time sequence data collection; extracting stable static structure entities according to the design data, and establishing an initial anchor point candidate area; after the time sequence point cloud data set coordinates are reconstructed according to the collection coordinates and collection parameters of the laser radar, loss identification under time sequence consistency loss network is performed; component semantic identification is performed on the time sequence point cloud data set, and a standardized component template is called; after the transformer substation scene is partitioned and fused, supplementary fusion is performed by using the standardized component template, and evaluation management is performed by using transformer substation digital modeling. The application can solve the technical problem of low precision of transformer substation scene digital modeling in the prior art, achieve the technical target of improving the precision of transformer substation scene digital modeling, and achieve the technical effect of improving the equipment recognition accuracy.

The application discloses a kind of virtual network card management methods based on KVM, applied to KVM virtualization environment, comprising: a, start kvm-br-sync daemon process, b, from virtual bridge module actual connection state of virtual bridge and virtual network card is obtained;C, the consistency of the mapping relationship and the actual connection state is verified;D, detect whether virtual machine is not running or virtual network card is invalid;E, detect whether the connection of virtual bridge and virtual network card is lost without reason;F, after network restart or virtual bridge / hardware network card fault repair, read the mapping relationship in database, detect whether the current virtual machine and virtual network card state match.Through database storage effective mapping, solve the core pain point of " unable to automatically recover after fault repair", guarantee virtual machine network reachability.