Transfer learning and arm compute Library

1 view (last 30 days)
Ahmed Bouzidi
Ahmed Bouzidi on 4 Dec 2019
Answered: Puru Kathuria on 25 Feb 2021
Hello, is it possible to train a network from matlab using the arm compute library on a raspberry pi to measure this library performance and influence?
if yes can someone help me to find the way to do that ?
Thank you very much

Answers (1)

Puru Kathuria
Puru Kathuria on 25 Feb 2021
These stpes might help your requirements:
  1. Train a neural network for the designated problem in MATLAB
  2. Save the network in MATLAB in a .mat file
  3. Set up the environment on Raspberry Pi by Installing Arm compute Library (Latest supported version). [reference]
  4. Write a MATLAB script that loads the network & performs prediction. Deploy the code on Raspberry Pi. Generate the executable for raspberry pi for this script.
  5. Test the network on Raspberry Pi and measure the performace by profiling each function. (Use PIL workflow for profiling) [reference]

Categories

Find more on Image Data Workflows in Help Center and File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!