Convolutional neural network operation method and device

Abstract

The invention discloses a convolutional neural network operation method and device, and the method comprises the steps: sequentially reading a first activation value, a second activation value and a third activation value of a convolution module from an activation value storage, using the first convolution kernel to convolve the first activation value in the temporary value memory to obtain a second activation value, using the second convolution kernel to convolve the second activation value to obtain a third activation value, using the third convolution kernel to convolve the third activationvalue to obtain a fourth activation value, and sequentially covering and writing the fourth activation value in the activation value memory; reading the first activation value and the fourth activation value from the activation value memory, superposing the first activation value and the fourth activation value in the temporary value memory based on linear correction to obtain a fifth activationvalue with a residual error, and overwriting the fifth activation value into the activation value memory. According to the invention, the video memory occupancy rate of the convolutional neural network in operation can be reduced, and the resource utilization rate and the operation speed are improved.

Description

technical field [0001] The present invention relates to the field of neural networks, and more specifically, to a method and device for operating a convolutional neural network. Background technique [0002] In recent years, deep learning networks have developed in a deeper and larger direction. The ResNet series of networks have proved that deeper networks can fit more data, but deep networks impose requirements on GPU memory. The maximum video memory supported by a modern GPU (graphics processing unit) is only 32G. Due to the high price of the GPU, expanding the number of GPUs will put a high pressure on the cost of the enterprise. Therefore, it is particularly important to save the GPU video memory occupied during the calculation process. [0003] ResNet50 is a 50-layer convolutional neural network that is widely used in image classification, object detection and other tasks in computer vision. ResNet50 is composed of multiple modules. In its forward calculation, the act...

AI Technical Summary

Problems solved by technology

ResNet50 is composed of multiple modules, and the activation value calculated in the middle will be saved in its forward calculation, which takes up a lot of video memory, and most of the activation value will not be used in the subsequent calculation process, thus causing a waste of resources

[0004] There is currently no effective solution to the problems of low reuse rate of intermediate activation values ​​of convolutional neural networks, occupying video memory, and slowing down the running speed in the prior art.

Images