A system for constructing a blood vessel lesion risk assessment model based on a large language model

Abstract

The application is based on a blood vessel lesion risk assessment model construction system of a large language model: a large model module uses a large language model to analyze model construction information and sends data storage instructions; a data storage module divides a target temporary storage area in a dynamic storage area and assigns a temporary storage address, stores model construction information to the target temporary storage area, and feeds back the temporary storage address to the large model module; the large model module sends data preprocessing instructions; the data preprocessing module preprocesses historical data and feeds back preprocessing operation completion information to the large model module; the large model module sends model training instructions; the model training module calls a target artificial intelligence model in a fixed storage area, calls preprocessed historical data and model configuration parameters in the temporary storage address for training, obtains optimal model parameters when the model converges, and feeds back convergence information to the large model module; the large model module sends model transfer instructions; and the data storage module transfers the optimal model and related information to the fixed storage area.

Images