Model accelerated training method and device based on training data similarity aggregation

Abstract

The invention discloses a model accelerated training method and device based on training data similarity aggregation, and the method comprises the steps: taking a part of minimized training data as a starting point, extracting data with poor prediction from a prediction result of a current model in a mode of random sampling and random increment in each round, and sampling additional training data in a clustering extraction mode, therefore, the most representative training information is obtained, and the training efficiency of each round is improved. The data set scale of each round of model training is reduced, the training time is greatly shortened, clustering does not need an accurate result, the number of iterations can be reduced or a faster and simpler clustering method is used, and the total training time of each round is still much shorter than that of original full training set training on the whole; the training data selected in each round is targeted, the images with inference errors are selected for training, the back propagation gradient can be obtained to the maximum extent, the probability of falling into the local optimal solution during training is reduced, dynamic adjustment in the training process is facilitated, and the optimal training result is achieved.

Description

technical field [0001] The invention relates to the technical field of image detection, in particular to a model acceleration training method and device based on similarity aggregation of training data. Background technique [0002] Image classification and image detection technologies based on deep learning are widely used technologies in the field of artificial intelligence. Different from traditional methods, image classification and image detection technology need to be extracted based on rules, and then trained for feature vectors. The image detection technology based on deep learning can automatically extract image features through a multi-layer convolutional neural network. The model receives the original pixel matrix input of the image and obtains end-to-end image recognition results. The deep neural network model is trained from a large number of pictures, which has better generalization ability, and the accuracy and anti-interference ability of image classification...

AI Technical Summary

Problems solved by technology

However, such an augmentation operation will add a lot of redundant and repeated information to the training data, which will slow down the efficiency of model training.

Images