kgubbi /  AI-design-contest-KWS-cnn

Created
Maintained by kgubbi
CNN Accelerator ProjectThe CNN Accelerator Project is a high-performance hardware solution designed to accelerate the execution of Convolutional Neural Networks (CNNs) on FPGA platforms. By leveraging custom-designed modules for weight management, input buffering, multiply-accumulate operations, and activation functions, this project significantly enhances the efficiency and speed of CNN computations. Integrated with a flexible state machine and robust control logic, the accelerator provides seamless interfacing with external systems via a comprehensive Wishbone bus and GPIO setup.Specifically optimized for efficient keyword spotting, our CNN Accelerator is ideal for applications in edge computing and AI-driven tasks, offering a powerful, scalable, and energy-efficient alternative to traditional software-based neural network processing.
Members 1
Kevin Immanuel Gubbi committed a month ago

AI-design-contest-KWS-cnn

CNN Accelerator Project The CNN Accelerator Project is a high-performance hardware solution designed to accelerate the execution of Convolutional Neural Networks (CNNs) on FPGA platforms. By leveraging custom-designed modules for weight management, input buffering, multiply-accumulate operations, and activation functions, this project significantly enhances the efficiency and speed of CNN computations. Integrated with a flexible state machine and robust control logic, the accelerator provides seamless interfacing with external systems via a comprehensive Wishbone bus and GPIO setup.

Specifically optimized for efficient keyword spotting, our CNN Accelerator is ideal for applications in edge computing and AI-driven tasks, offering a powerful, scalable, and energy-efficient alternative to traditional software-based neural network processing.


In order to interact with this project, make sure you have provided your ssh key here.

By convention, the original / primary remote repository is called origin.

Instructions

Local Project Setup - Git Commands

If you are starting a brand new project without a local repository, here are the suggested Git commands:

git clone ssh://git@repositories.efabless.com/kgubbi/AI-design-contest-KWS-cnn.git
cd AI-design-contest-KWS-cnn
... do some changes
git push -u origin main

If you already have a local Git repository, initialized with git init, you will need to associate your local repository with the Efabless Repository remote location. The following Git commands will add Efabless Repository as a remote repository and push your changes to the main branch.

git remote add origin ssh://git@repositories.efabless.com/kgubbi/AI-design-contest-KWS-cnn.git
git push -u origin main

If you already have a remote repository, for example one on GitHub, use the Git command below to add another remote Git repo (make sure that each repo has its unique ID, e.g. origin, ef-repo in the example below).

git remote add ef-repo ssh://git@repositories.efabless.com/kgubbi/AI-design-contest-KWS-cnn.git
git push -u ef-repo main

How To Update Your Project Cover Image

A default cover.png has already been added to the root directory of this remote repository. If your project is not private, this image can be viewed in the project search results list. To show your own cover image, replace cover.png with a file called 'cover' and an allowed image extension ('.jpg', '.jpeg', '.png', '.gif', '.webp') and then push your changes. Maximum allowed file size: 1 MB.