MSc thesis project proposal

Nest Loop Neural Network

Project outside the university

COMPAAN BV

Since the NN is written as a Nested Loop Program, we can convert the C-code into an implementation on a FPGA. For this purpose, we use the Compaan tool flow (www.compaandesign.com). For the synthesis of the mathematical functions used in the NN, we will use the Vivado Xilinx HLS tools. We want to show that the NN also works on a FPGA.

Assignment

Steps performed 1. Develop a testbench of a Neural Network in C-code that can detect handwritten letter using the MNIST database 2. Convert the C-code into an equivalent NN implementation on an FPGA 3. Proof that the obtained NN can detect the handwriting like the C-code 4. Determine the performance of the design and resource uses

Requirements

Required: Understanding of C-code and VHDL Understanding of FPGA flows Experience with Xilinx Vivado HLS tooling

Contact

dr.ir. René van Leuken

Signal Processing Systems Group

Department of Microelectronics

Last modified: 2021-12-06