Neural Network-Oriented Reconfigurable Pooling Operation Module Structure and Its Implementation Method

Abstract

The invention discloses a neural network-oriented reconfigurable pooling operation module structure, which includes a dynamic configuration module and a pooling calculation module. The pooling calculation module is used to complete the calculation process of the pooling layer. The role of the dynamic configuration module is Next, different pooling methods are implemented, and the calculation scale of the pooling operation is dynamically configured; the current input queue is read from the data memory after the last calculation, and the result is transmitted to the data memory through the output queue. The present invention also discloses a method for realizing a neural network-oriented reconfigurable pooling operation module structure. Through external control signals, each unit of the module is dynamically configured to control the pooling calculation module. According to the control signals sent by each unit, The calculation module performs corresponding calculation operations and gives accurate output results. This technical solution can not only reduce the complexity and time of the calculation process, but also reduce power consumption, and improve the flexibility of calculation and the utilization rate of units.

Description

technical field [0001] The invention belongs to the technical field of artificial neural networks, in particular to a reconfigurable pooling operation module structure oriented to neural networks. Background technique [0002] As the network becomes larger and larger, the neural network system needs to perform corresponding calculation operations on a large amount of input data, which will generate a large amount of power consumption. At present, a lot of research has been done on algorithms, software, and hardware. However, low energy consumption still faces new challenges. [0003] By compressing the feature map, the pooling operation can reduce the size of the feature map and simplify the computational complexity of the network; on the other hand, perform feature compression to extract main features. The pooling operation shows superiority in terms of performance and flexibility when dealing with neural networks. [0004] Since the pooling operation is an important proc...

AI Technical Summary

Problems solved by technology

However, low energy consumption still faces new challenges

Images