Model training method and device, storage medium and electronic device

Abstract

The invention discloses a model training method and device, a storage medium and an electronic device. The method comprises the steps that a training request is acquired, the training request is usedfor requesting to train a first neural network model, a second neural network model is obtained, and the second neural network model is used for identifying a picture with the picture type being a first type; a training set is determined through a third neural network model, the third neural network model is used for identifying the picture with the picture type being a target type, the target type comprises a first type, the training set comprises the picture which is identified from the first set and associated with the first type, and the picture in the first set is not marked with a firstidentifier; and the first neural network model is trained through the training set to obtain a second neural network model, the number of layers of the neural network in the second neural network model being different from the number of layers of the neural network in the second neural network model. The technical problem that the time cost for training the neural network model is high is solved.

Description

technical field [0001] The present invention relates to the field of the Internet, in particular to a model training method and device, a storage medium, and an electronic device. Background technique [0002] In machine learning and related fields, the computational model of artificial neural networks is inspired by the central nervous system of animals, which can rely on a large number of inputs and generally unknown approximate functions for estimation. Artificial neural networks are usually presented as interconnected "neurons", Because they are systems of adaptive nature, they can be used for machine learning and pattern recognition. [0003] For example, a neural network for handwriting recognition is defined by a set of input neurons that are likely to be activated by pixels of an input image. Backward weighted, and transformed by a function (determined by the designer of the network), the actuations of these neurons are recognized by other neurons and then passed on...

AI Technical Summary

Problems solved by technology

However, in related technologies, there is a high dependence on the labeling of training data information, which often requires large-scale training data with fully labeled information, and the cost of labeling is very high.

[0005] For the above problems, no effective solution has been proposed

Images