Convolution processing method and device, convolutional neural network accelerator and storage medium

The invention relates to the technical field of artificial intelligence, and provides a convolution processing method and device, a convolutional neural network accelerator and a storage medium. The method can improve the efficiency of convolution processing, and mainly comprises the steps of determining a plurality of data sets to be multiplied and added, wherein each data set comprises the weight data and the feature data; according to the number of the data sets, determining the number of assembly lines for multiplication and addition operations, wherein the number of the assembly lines and the number of the data sets have an integral multiple relation and are smaller than the number of the data sets; and equally dividing the plurality of data sets to each assembly line for multiplication and addition operations. 本申请涉及人工智能技术领域，提供一种卷积处理方法、装置、卷积神经网络加速器和存储介质，该方法可以提高卷积处理的效率，该方法主要包括：确定待乘加运算的多个数据组；各数据组均包括权重数据和特征数据；按照数据组的数量，确定用于乘加运算的流水线的数量；流水线的数量与数据组的数量具有整数倍关系，且小于数据组的数量；将多个数据组等份划分至各条流水线进行乘加运算。