Marking model training and marking method, device and system, storage medium and equipment

Abstract

The invention relates to a marking model training method, device and system, a marking method, device and system, a storage medium and equipment. The marking model training method comprises the steps of obtaining an initial model of a marking model; executing an iterative operation; the iterative operation comprises the steps of training an initial model based on an initial training sample to obtain a preprocessing model; inputting the current to-be-marked image into the preprocessing model, and outputting a current preprocessing marking result; taking the preprocessing model as a new initial model, taking the initial training sample and the current training sample as new initial training samples, returning to execute the iterative operation until a preset condition is met, and taking the preprocessing model meeting the preset condition as a marking model; wherein the current training sample is the current to-be-marked graph and the current preprocessing marking result conforming to the preset standard. By adopting the technical scheme provided by the invention, the generalization ability and / or robustness of the marking model can be improved without additionally increasing the marking cost of excessive training samples.

Description

technical field [0001] The present application relates to the technical field of artificial intelligence image automatic labeling, in particular to a labeling model training, labeling method, device, system, storage medium and equipment. Background technique [0002] With the improvement of the level of science and technology, the marking technology based on artificial intelligence has been developed rapidly and widely used. [0003] In the process of training the marked model based on certain methods (such as: supervised learning, semi-supervised learning), it needs to rely on training samples and the marks of the training samples to continuously optimize the training of the model. In the prior art, in order to improve the generalization ability and / or robustness of the labeling model, a large number of training samples are often required, which causes the labeling of the training samples to take a lot of time; in addition, the training time of the model is too long, Moreo...

AI Technical Summary

Problems solved by technology

In the prior art, in order to improve the generalization ability and / or robustness of the labeling model, a large number of training samples are often required, which causes the labeling of the training samples to take a lot of time; in addition, the training time of the model is too long, And because the training data is huge, the system hardware requirements are high, resulting in high cost of the system

Images