5 results about "Host model" patented technology

A system for accessing computing elements of a cloud computing infrastructure includes a dynamic host interface associated with an identity-based access management system. The dynamic host interface is configured to receive host catalogs from components of the cloud computing infrastructure, each host catalog including one or more host sets, each of the one or more host sets including references to one or more hosts, where each of the one or more hosts represents an addressable computing element. The system further includes a plugin-based hosts model in a memory, and being configured to receive, from the dynamic host interface, the host catalogs for identity-based access to the one or more host sets by a user of a client computer to address the addressable computing element represented by the host.

The invention relates to the technical field of network security operation and maintenance, and discloses an SCCD-plus model-based network security equipment unified digital governance method, which comprises the following steps of: constructing information following IEC 61850 system standards, expanding asset ledgers, network topology and the like, redefining SCCD model files of network security equipment and a station control layer host model, and establishing an SCCD model file of the network security equipment and the station control layer host model; the method comprises the following steps: acquiring actual operation configuration values of network security equipment and a station level host, checking and merging with a configuration fixed value baseline in an SCCD model file, generating an SCCD-plus archived file, instantiating a model of target equipment, generating a CID-plus file, and applying the CID-plus file to complete automatic configuration. According to the invention, unified acquisition and management of incremental equipment and stock equipment are realized, automatic configuration, compliance check and security policy deployment of network security equipment are supported, the efficiency and accuracy of network security management are effectively improved, and the operation and maintenance cost is reduced.

This invention provides a temporal model representation method based on task arithmetic and weight alignment, comprising: acquiring meteorological monitoring time-series data; preprocessing the meteorological monitoring time-series data to obtain processed meteorological time-series data; inputting the processed meteorological time-series data into a target host model for parameter fusion for prediction processing to obtain meteorological prediction results; the target host model for parameter fusion is a network model after covering the embedding layer of the target host model based on an enhanced embedding module; the enhanced embedding module is a feature processor based on a weighted linear interpolation method, which weights and fuses the parameters of the source auxiliary model and the target host model within the embedding layer; both the target host model and the source auxiliary model are pre-trained temporal models. This significantly reduces the deployment and operational overhead of meteorological forecasting systems, achieving high efficiency, high accuracy, and strong generalization ability in accurate prediction of meteorological elements.

The invention relates to the field of data processing, and discloses a machine learning model management method and device based on MLOps, equipment and a storage medium, which are used for improving the management efficiency of a model. The MLOps-based machine learning model management method comprises the steps of obtaining target original data and model related information submitted by a user, registering a model to a trusteeship platform according to the model related information, and associating a code version with a data version to obtain a trusteeship model; dividing the target original data into a training data set and a test data set; triggering a corresponding training task based on the training data set, the test data set and model version information corresponding to the hosting model; executing a model training task in a specified environment, and after training is completed, evaluating the model to generate a structured training report; and when a preset iteration triggering condition is met, performing iteration processing and updating the hosting model.

Model server replicas are initialized on a set of first host machines. The model server replicas are each configured to execute an instance of a machine-learned model by obtaining first model image partitions. Each model image partition stores a separate portion of the model. Initializer nodes are executed on a set of second host machines that are selected based on a geographic location of the set of first host machines. Each of the initializer nodes comprises a local image registry mirror provisioned with the model image partitions. Each of the model server replicas are configured such that the model server replica pulls the model image partitions from the local image registry mirror of an initializer node.