Skip to content

Latest commit

 

History

History
executable file
·
77 lines (50 loc) · 1.9 KB

File metadata and controls

executable file
·
77 lines (50 loc) · 1.9 KB

Introduction

The SISLAB team in LSI Design Contest 2018.

Design challenge: "Neural Network (Backpropagation)"

The goals of project is learning about neural network, the flow from algorithm to hardware architecture, having an result.

Team members

  • Huy-Hung Ho
  • Van-Thuat Nguyen
  • Xuan-Thuan Nguyen
  • Van-Dung Nguyen

Architecture

Phases

  1. Study about Supervised learning of Neural Network (using Backpropagation) algorithm (find out at less a optimization).
  2. Software simulation by C/C++ (or an other) language.
  3. Hardware simulation by VHDL language.
  4. Synthesis in FPGA. (if enough time)

Plan

8th Oct - 7th Nov: (31 days)

  • Studying algorithm
    • Collecting the state-of-the-art papers (keyword: hardware backpropagation neural network)
    • Try to find the optimization
  • Sofware simulation (matlab, C/C++, speeding Matlab code)

8th Nov - 10th Dec: (34 days)

  • Trainning VHDL tools (ubuntu, vsim, git)
  • Studying FPGA synthesis and Synopsys tools (Synphony Model™ Compiler, Synplify Pro, Design Compiler)
  • Architecture demo of ANN, writing VHDL code a part
  • Writing VHDL for full ANN

11th Doc - 7th Jan: (28 days)

  • Writing VHDL

8th Jan - 30th Jan: (23 days)

  • Verification, testing in FPGA

Timeline

  • Start time : 8th Oct, 2017
  • Deadline : 31st Jan, 2018
  • Timing : 116 days
  • Do the exam : 11th Dec - 7th Jan

Service