How to Build a Neural Network on an FPGA

In this tutorial, join Ari Mahpour as he explores the fascinating task of deploying neural networks on the PYNQ-Z2 FPGA board. Learn step-by-step how to train an iris classification model using the powerful hls4ml framework and validate it against both local and FPGA models.

**What You Will Learn:**
- Setting up your PYNQ-Z2 board for AI deployment
- Training and optimizing neural networks for FPGA
- Leveraging the hls4ml repository for seamless FPGA integration
- Comparing performance between CPU/GPU and FPGA models

**Resources Mentioned:**

✨ Don't Miss Out:

#NeuralNetwork #FPGA #ai

0:00 Intro
0:43 A Note before We Begin
3:17 Dataset Overview
8:04 Building the Model & Flash File
21:18 Running & Validating the Model
31:59 Wrapping Up

[robertoguerra]

Great video.

This task is a large chunk to byte for the fpga community… the intermediate steps would be DSPs, correlations, probability estimations and linear fittings (in multiple dimensions) in FPGA… then the engineering community would be ready to make a learning NN in FPGA.

It is not as simple as converting the c code to vhdl

[pardhuIOT]

Hello Sir,
"In the new version of Xilinx, Vivado HLS has been removed, but Vivado Accelerator is still trying to access Vivado HLS, resulting in the error: 'Vivado HLS installation not found.'
how to resolve this ?

[rafaasanchez]

Hello, I just send you an email. Great video!