Repository files navigation

This repository contains an implementations of a binary ResNet50FINN-style dataflow accelerator targeting Alveo boards.It is intended as a showcase of achievable throughput and latency for ImageNet clasiffication on FPGA, using dataflow execution and on-chip weight storage.

The repository is organized as follows:

• src: contains source code and submodules
  • hls: HLS custom building blocks and submodules to FINN librares (FINN andFINN-HLSLib)
  • w1a2-v1.0: pre-build weights, thresholds, directives and configuration files for Binary ResNet50
• compile: contains scripts for accelerator compilation (Vivado HLS CSynth + Vivado Synthesis)
• link: contains scripts for accelerator linking into the Alveo platform with Vitis
• host: python and Jupyter host code, using PYNQ for Alveo

The Accelerator is built usingVitis 2019.2. We recommend using this version,otherwise changes might be required to source and/or Makefiles for things towork.

To build the accelerator, clone the repository (using--recursive to pull submodules), after which:

cd ResNet50-PYNQ/compilemake NET=w1a2_v1.0cd ../linkmake

See the specificCompile andLink documentation for further info.

After you have built the accelerator, you caninstall the required files inthehost folder. Move in the cloned repo and domake install

cd ResNet50-PYNQmake install

You can then run the included Jupyter notebook or the Python multithreaded inference example. If youwant to use the distributed PYNQ python package, please read below.If you want to run example Python inference code, please see thehost code documentation.

Install theresnet50-pynq package usingpip:

pip install resnet50-pynq

After the package is installed, to get your own copy of the available notebooksrun:

pynq get-notebooks ResNet50

You can then try things out by doing:

cd pynq-notebooksjupyter notebook

There are a number of additional options for thepynq get-notebooks command,you can list them by typing

pynq get-notebooks --help

You can also refer to the officialPYNQ documentation for more informationregarding thePYNQ Command Line Interface and in particular theget-notebooks command.

Currently, we distribute the overlay only for the following Alveo boards andshells:

Designs are built usingVitis 2019.2.

Lucian Petrica @ Xilinx Research Labs.