Využití vysokoúrovňové syntézy pro urychlení vývoje aplikací na programovatelné logice

Abstract

The first part of the bachelor thesis deals with the creation of a neural network with feedback that can recognize the handwritten data set of MNIST numbers. In the second part of the bachelor’s thesis, I deal with the implementation of a neural network on an FPGA device using a compiler, where I monitor changes in the speed of code execution based on changes before translation.

Description

Subject(s)

Neural network, FPGA, MNIST, compiler, neural network with backpropagation, OpenCL, C ++

Citation